This file documents NCO, a collection of utilities to manipulate and analyze netCDF files.
Copyright © 1995–2009 Charlie Zender
This is the first edition of the NCO User's Guide,
and is consistent with version 2 of texinfo.tex.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.3 or any later version published by the Free Software Foundation; with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. The license is available online at http://www.gnu.org/copyleft/fdl.html
The original author of this software, Charlie Zender, wants to improve it
with the help of your suggestions, improvements, bug-reports, and patches.
Charlie Zender <surname at uci dot edu> (yes, my surname is zender)
3200 Croul Hall
Department of Earth System Science
University of California, Irvine
Irvine, CA 92697-3100
|
Note to readers of the NCO User's Guide in HTML format:
The NCO User's Guide in PDF format
(also on SourceForge)
contains the complete NCO documentation.
|
The netCDF Operators, or NCO, are a suite of programs known as operators. The operators facilitate manipulation and analysis of data stored in the self-describing netCDF format, available from (http://www.unidata.ucar.edu/packages/netcdf). Each NCO operator (e.g., ncks) takes netCDF input file(s), performs an operation (e.g., averaging, hyperslabbing, or renaming), and outputs a processed netCDF file. Although most users of netCDF data are involved in scientific research, these data formats, and thus NCO, are generic and are equally useful in fields from agriculture to zoology. The NCO User's Guide illustrates NCO use with examples from the field of climate modeling and analysis. The NCO homepage is http://nco.sf.net, and there is a mirror at http://dust.ess.uci.edu/nco.
This documentation is for NCO version 4.0.0. It was last updated 15 July 2009. Corrections, additions, and rewrites of this documentation are very welcome.
Enjoy,
Charlie Zender
NCO is the result of software needs that arose while I worked
on projects funded by NCAR, NASA, and ARM.
Thinking they might prove useful as tools or templates to others,
it is my pleasure to provide them freely to the scientific community.
Many users (most of whom I have never met) have encouraged the
development of NCO.
Thanks espcially to Jan Polcher, Keith Lindsay, Arlindo da Silva,
John Sheldon, and William Weibel for stimulating suggestions and
correspondence.
Your encouragment motivated me to complete the NCO User's Guide.
So if you like NCO, send me a note!
I should mention that NCO is not connected to or
officially endorsed by Unidata, ACD, ASP,
CGD, or Nike.
Charlie Zender
Major feature improvements entitle me to write another Foreword. In the last five years a lot of work has been done to refine NCO. NCO is now an open source project and appears to be much healthier for it. The list of illustrious institutions that do not endorse NCO continues to grow, and now includes UCI.
Charlie Zender
The most remarkable advances in NCO capabilities in the last few years are due to contributions from the Open Source community. Especially noteworthy are the contributions of Henry Butowsky and Rorik Peterson.
Charlie Zender
NCO has been generously supported from 2004–2008 by US National Science Foundation (NSF) grant IIS-0431203. This support allowed me to maintain and extend core NCO code, and others to advance NCO in new directions: Gayathri Venkitachalam helped implement MPI; Harry Mangalam improved regression testing and benchmarking; Daniel Wang developed the server-side capability, SWAMP; and Henry Butowsky, a long-time contributor, developed ncap2. This support also led NCO to debut in professional journals and meetings. The personal and professional contacts made during this evolution have been immensely rewarding.
Charlie Zender
This manual describes NCO, which stands for netCDF Operators. NCO is a suite of programs known as operators. Each operator is a standalone, command line program executed at the shell-level like, e.g., ls or mkdir. The operators take netCDF files (including HDF5 files constructed using the netCDF API) as input, perform an operation (e.g., averaging or hyperslabbing), and produce a netCDF file as output. The operators are primarily designed to aid manipulation and analysis of data. The examples in this documentation are typical applications of the operators for processing climate model output. This stems from their origin, though the operators are as general as netCDF itself.
The complete NCO source distribution is currently distributed
as a compressed tarfile from
http://sf.net/projects/nco
and from
http://dust.ess.uci.edu/nco/nco.tar.gz.
The compressed tarfile must be uncompressed and untarred before building
NCO.
Uncompress the file with ‘gunzip nco.tar.gz’.
Extract the source files from the resulting tarfile with ‘tar -xvf
nco.tar’.
GNU tar lets you perform both operations in one step
with ‘tar -xvzf nco.tar.gz’.
The documentation for NCO is called the NCO User's Guide. The User's Guide is available in Postscript, HTML, DVI, TeXinfo, and Info formats. These formats are included in the source distribution in the files nco.ps, nco.html, nco.dvi, nco.texi, and nco.info*, respectively. All the documentation descends from a single source file, nco.texi 1. Hence the documentation in every format is very similar. However, some of the complex mathematical expressions needed to describe ncwa can only be displayed in DVI, Postscript, and PDF formats.
A complete list of papers and publications on/about NCO is available on the NCO homepage. Most of these are freely available. The primary refereed publications are fxm ZeM06 and fxm Zen07. These contain copyright restrictions which limit their redistribution, but they are freely available in preprint form from the NCO.
If you want to quickly see what the latest improvements in NCO are (without downloading the entire source distribution), visit the NCO homepage at http://nco.sf.net. The HTML version of the User's Guide is also available online through the World Wide Web at URL http://nco.sf.net/nco.html. To build and use NCO, you must have netCDF installed. The netCDF homepage is http://www.unidata.ucar.edu/packages/netcdf.
New NCO releases are announced on the netCDF list
and on the nco-announce mailing list
http://lists.sf.net/mailman/listinfo/nco-announce.
NCO has been successfully ported and tested and is known to work on the following 32- and 64-bit platforms: IBM AIX 4.x, 5.x, FreeBSD 4.x, GNU/Linux 2.x, LinuxPPC, LinuxAlpha, LinuxARM, LinuxSparc64, SGI IRIX 5.x and 6.x, MacOS X 10.x, NEC Super-UX 10.x, DEC OSF, Sun SunOS 4.1.x, Solaris 2.x, Cray UNICOS 8.x–10.x, and MS Windows95 and all later versions. If you port the code to a new operating system, please send me a note and any patches you required.
The major prerequisite for installing NCO on a particular platform is the successful, prior installation of the netCDF library (and, as of 2003, the UDUnits library). Unidata has shown a commitment to maintaining netCDF and UDUnits on all popular UNIX platforms, and is moving towards full support for the Microsoft Windows operating system (OS). Given this, the only difficulty in implementing NCO on a particular platform is standardization of various C and Fortran interface and system calls. NCO code is tested for ANSI compliance by compiling with C compilers including those from GNU (‘gcc -std=c99 -pedantic -D_BSD_SOURCE -D_POSIX_SOURCE’ -Wall) 2, Comeau Computing (‘como --c99’), Cray (‘cc’), HP/Compaq/DEC (‘cc’), IBM (‘xlc -c -qlanglvl=extc99’), Intel (‘icc -std=c99’), NEC (‘cc’), PathScale (QLogic) (‘pathcc -std=c99’), PGI (‘pgcc -c9x’), SGI (‘cc -c99’), and Sun (‘cc’). NCO (all commands and the libnco library) and the C++ interface to netCDF (called libnco_c++) comply with the ISO C++ standards as implemented by Comeau Computing (‘como’), Cray (‘CC’), GNU (‘g++ -Wall’), HP/Compaq/DEC (‘cxx’), IBM (‘xlC’), Intel (‘icc’), NEC (‘c++’), PathScale (Qlogic) (‘pathCC’), PGI (‘pgCC’), SGI (‘CC -LANG:std’), and Sun (‘CC -LANG:std’). See nco/bld/Makefile and nco/src/nco_c++/Makefile.old for more details and exact settings.
Until recently (and not even yet), ANSI-compliant has meant
compliance with the 1989 ISO C-standard, usually called C89 (with
minor revisions made in 1994 and 1995).
C89 lacks variable-size arrays, restricted pointers, some useful
printf formats, and many mathematical special functions.
These are valuable features of C99, the 1999 ISO C-standard.
NCO is C99-compliant where possible and C89-compliant where
necessary.
Certain branches in the code are required to satisfy the native
SGI and SunOS C compilers, which are strictly ANSI
C89 compliant, and cannot benefit from C99 features.
However, C99 features are fully supported by modern AIX,
GNU, Intel, NEC, Solaris, and UNICOS
compilers.
NCO requires a C99-compliant compiler as of NCO
version 2.9.8, released in August, 2004.
The most time-intensive portion of NCO execution is spent in
arithmetic operations, e.g., multiplication, averaging, subtraction.
These operations were performed in Fortran by default until August,
1999.
This was a design decision based on the relative speed of Fortran-based
object code vs. C-based object code in late 1994.
C compiler vectorization capabilities have dramatically improved
since 1994.
We have accordingly replaced all Fortran subroutines with C functions.
This greatly simplifies the task of building NCO on nominally
unsupported platforms.
As of August 1999, NCO built entirely in C by default.
This allowed NCO to compile on any machine with an
ANSI C compiler.
In August 2004, the first C99 feature, the restrict type
qualifier, entered NCO in version 2.9.8.
C compilers can obtain better performance with C99 restricted
pointers since they inform the compiler when it may make Fortran-like
assumptions regarding pointer contents alteration.
Subsequently, NCO requires a C99 compiler to build correctly
3.
In January 2009, NCO version 3.9.6 was the first to link to the GNU Scientific Library (GSL). GSL must be version 1.4 or later. NCO, in particular ncap2, uses the GSL special function library to evaluate geoscience-relevant mathematics such as Bessel functions, Legendre polynomials, and incomplete gamma functions (see GSL special functions).
In June 2005, NCO version 3.0.1 began to take advantage
of C99 mathematical special functions.
These include the standarized gamma function (called tgamma()
for “true gamma”).
NCO automagically takes advantage of some GNU
Compiler Collection (GCC) extensions to ANSI C.
As of July 2000 and NCO version 1.2, NCO no
longer performs arithmetic operations in Fortran.
We decided to sacrifice executable speed for code maintainability.
Since no objective statistics were ever performed to quantify
the difference in speed between the Fortran and C code,
the performance penalty incurred by this decision is unknown.
Supporting Fortran involves maintaining two sets of routines for every
arithmetic operation.
The USE_FORTRAN_ARITHMETIC flag is still retained in the
Makefile.
The file containing the Fortran code, nco_fortran.F, has been
deprecated but a volunteer (Dr. Frankenstein?) could resurrect it.
If you would like to volunteer to maintain nco_fortran.F please
contact me.
NCO has been successfully ported and tested on the Microsoft
Windows (95/98/NT/2000/XP) operating systems.
The switches necessary to accomplish this are included in the standard
distribution of NCO.
Using the freely available Cygwin (formerly gnu-win32) development
environment
4, the compilation process is very similar to
installing NCO on a UNIX system.
Set the PVM_ARCH preprocessor token to WIN32.
Note that defining WIN32 has the side effect of disabling
Internet features of NCO (see below).
NCO should now build like it does on UNIX.
The least portable section of the code is the use of standard
UNIX and Internet protocols (e.g., ftp, rcp,
scp, sftp, getuid, gethostname, and header
files <arpa/nameser.h> and
<resolv.h>).
Fortunately, these UNIX-y calls are only invoked by the single
NCO subroutine which is responsible for retrieving files
stored on remote systems (see Remote storage).
In order to support NCO on the Microsoft Windows platforms,
this single feature was disabled (on Windows OS only).
This was required by Cygwin 18.x—newer versions of Cygwin may
support these protocols (let me know if this is the case).
The NCO operators should behave identically on Windows and
UNIX platforms in all other respects.
Like all executables, the NCO operators can be built using dynamic linking. This reduces the size of the executable and can result in significant performance enhancements on multiuser systems. Unfortunately, if your library search path (usually the LD_LIBRARY_PATH environment variable) is not set correctly, or if the system libraries have been moved, renamed, or deleted since NCO was installed, it is possible NCO operators will fail with a message that they cannot find a dynamically loaded (aka shared object or ‘.so’) library. This will produce a distinctive error message, such as ‘ld.so.1: /usr/local/bin/ncea: fatal: libsunmath.so.1: can't open file: errno=2’. If you received an error message like this, ask your system administrator to diagnose whether the library is truly missing 5, or whether you simply need to alter your library search path. As a final remedy, you may re-compile and install NCO with all operators statically linked.
netCDF version 2 was released in 1993.
NCO (specifically ncks) began soon after this in 1994.
netCDF 3.0 was released in 1996, and we were eager to reap the
performance advantages of the newer netCDF implementation.
One netCDF3 interface call (nc_inq_libvers) was added to
NCO in January, 1998, to aid in maintainance and debugging.
In March, 2001, the final conversion of NCO to netCDF3
was completed (coincidentally on the same day netCDF 3.5 was
released).
NCO versions 2.0 and higher are built with the
-DNO_NETCDF_2 flag to ensure no netCDF2 interface calls
are used.
However, the ability to compile NCO with only netCDF2
calls is worth maintaining because HDF version 4
6
(available from HDF)
supports only the netCDF2 library calls
(see http://hdf.ncsa.uiuc.edu/UG41r3_html/SDS_SD.fm12.html#47784).
Note that there are multiple versions of HDF.
Currently HDF version 4.x supports netCDF2 and thus
NCO version 1.2.x.
If NCO version 1.2.x (or earlier) is built with only
netCDF2 calls then all NCO operators should work with
HDF4 files as well as netCDF files
7.
The preprocessor token NETCDF2_ONLY exists
in NCO version 1.2.x to eliminate all netCDF3
calls.
Only versions of NCO numbered 1.2.x and earlier have this
capability.
The NCO 1.2.x branch will be maintained with bugfixes only
(no new features) until HDF begins to fully support the
netCDF3 interface (which is employed by NCO 2.x).
If, at compilation time, NETCDF2_ONLY is defined, then
NCO version 1.2.x will not use any netCDF3 calls
and, if linked properly, the resulting NCO operators will work
with HDF4 files.
The Makefile supplied with NCO 1.2.x is written
to simplify building in this HDF capability.
When NCO is built with make HDF4=Y, the Makefile
sets all required preprocessor flags and library links to build
with the HDF4 libraries (which are assumed to reside under
/usr/local/hdf4, edit the Makefile to suit your
installation).
HDF version 5 became available in 1999, but did not support netCDF (or, for that matter, Fortran) as of December 1999. By early 2001, HDF5 did support Fortran90. In 2004, Unidata and NCSA began a project to implement the HDF5 features necessary to support the netCDF API. NCO version 3.0.3 added support for reading/writing netCDF4-formatted HDF5 files in October, 2005. See Selecting Output File Format for more details.
HDF support for netCDF was completed with HDF5 version version 1.8 in 2007. The netCDF front-end that uses this HDF5 back-end was completed and released soon after as netCDF version 4. Download it from the netCDF4 website.
NCO version 3.9.0 added support for all netCDF4 atomic data
types except NC_STRING in May, 2007.
Support for additional netCDF4 features has been incremental.
We add one netCDF4 feature at a time.
You must build NCO with netCDF4 to obtain this support.
The main netCDF4 features that NCO currently supports are the new
atomic data types and Lempel-Ziv compression.
The new atomic data types are NC_UBYTE, NC_USHORT,
NC_UINT, NC_INT64, and NC_UINT64.
Eight-byte integer support is an especially useful improvement from
netCDF3.
All NCO operators support these types, e.g., ncks
copies and prints them, ncra averages them, and
ncap2 processes algebraic scripts with them.
ncks prints compression information, if any, to screen.
NCO version 3.9.9 (June, 2009) added support for the
NC_STRING netCDF4 atomic data type.
Ragged arrays of strings are supported.
NCO version 3.9.1 (June, 2007) added support for netCDF4 Lempel-Ziv deflation. Lempel-Ziv deflation is a lossless compression technique. See Deflation for more details.
NCO version 3.9.9 (June, 2009) added support for netCDF4 chunking. See Chunking for more details.
netCDF4-enabled NCO handles netCDF3 files without change. In addition, it automagically handles netCDF4 (HDF5) files: If you feed NCO netCDF3 files, it produces netCDF3 output. If you feed NCO netCDF4 files, it produces netCDF4 output. Use the handy-dandy ‘-4’ switch to request netCDF4 output from netCDF3 input, i.e., to convert netCDF3 to netCDF4. See Selecting Output File Format for more details.
Use appropriate caution while netCDF4 is beta software. Problems with netCDF4 and HDF libraries are still being fixed. NCO support for netCDF4 atomic types is relatively untested. Binary NCO distributions (RPMs and debs) still use netCDF3.
For now you must build NCO from source to get netCDF4 support.
Typically, one specifies the root of the netCDF4-beta
installation directory. Do this with the NETCDF4_ROOT variable.
Then use your preferred NCO build mechanism, e.g.,
export NETCDF4_ROOT=/usr/local/netcdf4 # Set netCDF4 location
cd ~/nco;./configure --enable-netcdf4 # Configure mechanism -or-
cd ~/nco/bld;./make NETCDF4=Y allinone # Old Makefile mechanism
Our short term goal is to track the netCDF4-beta releases, keep the new netCDF4 atomic type support working, and iron out any problems. Our long term goal is to utilize more of the extensive new netCDF4 feature set. The next major netCDF4 feature we are likely to utilize is parallel I/O. We will enable this in the MPI netCDF operators.
We generally receive three categories of mail from users: help requests, bug reports, and feature requests. Notes saying the equivalent of "Hey, NCO continues to work great and it saves me more time everyday than it took to write this note" are a distant fourth.
There is a different protocol for each type of request. The preferred etiquette for all communications is via NCO Project Forums. Do not contact project members via personal e-mail unless your request comes with money or you have damaging information about our personal lives. Please use the Forums—they preserve a record of the questions and answers so that others can learn from our exchange. Also, since NCO is government-funded, this record helps us provide program officers with information they need to evaluate our project.
Before posting to the NCO forums described below, you might first register your name and email address with SourceForge.net or else all of your postings will be attributed to "nobody". Once registered you may choose to "monitor" any forum and to receive (or not) email when there are any postings including responses to your questions. We usually reply to the forum message, not to the original poster.
If you want us to include a new feature in NCO, check first to see if that feature is already on the TODO list. If it is, why not implement that feature yourself and send us the patch? If the feature is not yet on the list, then send a note to the NCO Discussion forum.
Read the manual before reporting a bug or posting a help request. Sending questions whose answers are not in the manual is the best way to motivate us to write more documentation. We would also like to accentuate the contrapositive of this statement. If you think you have found a real bug the most helpful thing you can do is simplify the problem to a manageable size and then report it. The first thing to do is to make sure you are running the latest publicly released version of NCO.
Once you have read the manual, if you are still unable to get NCO to perform a documented function, submit a help request. Follow the same procedure as described below for reporting bugs (after all, it might be a bug). That is, describe what you are trying to do, and include the complete commands (run with ‘-D 5’), error messages, and version of NCO (with ‘-r’). Post your help request to the NCO Help forum.
If you think you used the right command when NCO misbehaves, then you might have found a bug. Incorrect numerical answers are the highest priority. We usually fix those within one or two days. Core dumps and sementation violations receive lower priority. They are always fixed, eventually.
How do you simplify a problem that reveal a bug? Cut out extraneous variables, dimensions, and metadata from the offending files and re-run the command until it no longer breaks. Then back up one step and report the problem. Usually the file(s) will be very small, i.e., one variable with one or two small dimensions ought to suffice. Run the operator with ‘-r’ and then run the command with ‘-D 5’ to increase the verbosity of the debugging output. It is very important that your report contain the exact error messages and compile-time environment. Include a copy of your sample input file, or place one on a publically accessible location, of the file(s). Post the full bug report to the NCO Project buglist.
Build failures count as bugs.
Our limited machine access means we cannot fix all build failures.
The information we need to diagnose, and often fix, build failures
are the three files output by GNU build tools,
nco.config.log.${GNU_TRP}.foo,
nco.configure.${GNU_TRP}.foo,
and nco.make.${GNU_TRP}.foo.
The file configure.eg shows how to produce these files.
Here ${GNU_TRP} is the "GNU architecture triplet",
the chip-vendor-OS string returned by config.guess.
Please send us your improvements to the examples supplied in
configure.eg.
The regressions archive at http://dust.ess.uci.edu/nco/rgr
contains the build output from our standard test systems.
You may find you can solve the build problem yourself by examining the
differences between these files and your own.
The main design goal is command line operators which perform useful, scriptable operations on netCDF files. Many scientists work with models and observations which produce too much data to analyze in tabular format. Thus, it is often natural to reduce and massage this raw or primary level data into summary, or second level data, e.g., temporal or spatial averages. These second level data may become the inputs to graphical and statistical packages, and are often more suitable for archival and dissemination to the scientific community. NCO performs a suite of operations useful in manipulating data from the primary to the second level state. Higher level interpretive languages (e.g., IDL, Yorick, Matlab, NCL, Perl, Python), and lower level compiled languages (e.g., C, Fortran) can always perform any task performed by NCO, but often with more overhead. NCO, on the other hand, is limited to a much smaller set of arithmetic and metadata operations than these full blown languages.
Another goal has been to implement enough command line switches so that frequently used sequences of these operators can be executed from a shell script or batch file. Finally, NCO was written to consume the absolute minimum amount of system memory required to perform a given job. The arithmetic operators are extremely efficient; their exact memory usage is detailed in Memory Requirements.
NCO was developed at NCAR to aid analysis and manipulation of datasets produced by General Circulation Models (GCMs). Datasets produced by GCMs share many features with all gridded scientific datasets and so provide a useful paradigm for the explication of the NCO operator set. Examples in this manual use a GCM paradigm because latitude, longitude, time, temperature and other fields related to our natural environment are as easy to visualize for the layman as the expert.
NCO operators are designed to be reasonably fault tolerant, so
that if there is a system failure or the user aborts the operation (e.g.,
with C-c), then no data are lost.
The user-specified output-file is only created upon successful
completion of the operation
8.
This is accomplished by performing all operations in a temporary copy
of output-file.
The name of the temporary output file is constructed by appending
.pid<process ID>.<operator name>.tmp to the
user-specified output-file name.
When the operator completes its task with no fatal errors, the temporary
output file is moved to the user-specified output-file.
Note the construction of a temporary output file uses more disk space
than just overwriting existing files “in place” (because there may be
two copies of the same file on disk until the NCO operation
successfully concludes and the temporary output file overwrites the
existing output-file).
Also, note this feature increases the execution time of the operator
by approximately the time it takes to copy the output-file.
Finally, note this feature allows the output-file to be the same
as the input-file without any danger of “overlap”.
Other safeguards exist to protect the user from inadvertently overwriting data. If the output-file specified for a command is a pre-existing file, then the operator will prompt the user whether to overwrite (erase) the existing output-file, attempt to append to it, or abort the operation. However, in processing large amounts of data, too many interactive questions slows productivity. Therefore NCO also implements two ways to override its own safety features, the ‘-O’ and ‘-A’ switches. Specifying ‘-O’ tells the operator to overwrite any existing output-file without prompting the user interactively. Specifying ‘-A’ tells the operator to attempt to append to any existing output-file without prompting the user interactively. These switches are useful in batch environments because they suppress interactive keyboard input.
Adding variables from one file to another is often desirable. This is referred to as appending, although some prefer the terminology merging 9 or pasting. Appending is often confused with what NCO calls concatenation. In NCO, concatenation refers to splicing a variable along the record dimension. Appending, on the other hand, refers to adding variables from one file to another 10. In this sense, ncks can append variables from one file to another file. This capability is invoked by naming two files on the command line, input-file and output-file. When output-file already exists, the user is prompted whether to overwrite, append/replace, or exit from the command. Selecting overwrite tells the operator to erase the existing output-file and replace it with the results of the operation. Selecting exit causes the operator to exit—the output-file will not be touched in this case. Selecting append/replace causes the operator to attempt to place the results of the operation in the existing output-file, See ncks netCDF Kitchen Sink.
The simplest way to create the union of two files is
ncks -A fl_1.nc fl_2.nc
This puts the contents of fl_1.nc into fl_2.nc. The ‘-A’ is optional. On output, fl_2.nc is the union of the input files, regardless of whether they share dimensions and variables, or are completely disjoint. The append fails if the input files have differently named record dimensions (since netCDF supports only one), or have dimensions of the same name but different sizes.
Users comfortable with NCO semantics may find it easier to perform some simple mathematical operations in NCO rather than higher level languages. ncbo (see ncbo netCDF Binary Operator) does file addition, subtraction, multiplication, division, and broadcasting. ncflint (see ncflint netCDF File Interpolator) does file addition, subtraction, multiplication and interpolation. Sequences of these commands can accomplish simple but powerful operations from the command line.
The most frequently used operators of NCO are probably the averagers and concatenators. Because there are so many permutations of averaging (e.g., across files, within a file, over the record dimension, over other dimensions, with or without weights and masks) and of concatenating (across files, along the record dimension, along other dimensions), there are currently no fewer than five operators which tackle these two purposes: ncra, ncea, ncwa, ncrcat, and ncecat. These operators do share many capabilities 11, but each has its unique specialty. Two of these operators, ncrcat and ncecat, are for concatenating hyperslabs across files. The other two operators, ncra and ncea, are for averaging hyperslabs across files 12. First, let's describe the concatenators, then the averagers.
Joining independent files together along a record dimension is called
concatenation.
ncrcat is designed for concatenating record variables, while
ncecat is designed for concatenating fixed length variables.
Consider five files, 85.nc, 86.nc,
... 89.nc each containing a year's worth of data.
Say you wish to create from them a single file, 8589.nc
containing all the data, i.e., spanning all five years.
If the annual files make use of the same record variable, then
ncrcat will do the job nicely with, e.g.,
ncrcat 8?.nc 8589.nc.
The number of records in the input files is arbitrary and can vary from
file to file.
See ncrcat netCDF Record Concatenator, for a complete description of
ncrcat.
However, suppose the annual files have no record variable, and thus
their data are all fixed length.
For example, the files may not be conceptually sequential, but rather
members of the same group, or ensemble.
Members of an ensemble may have no reason to contain a record dimension.
ncecat will create a new record dimension (named record
by default) with which to glue together the individual files into the
single ensemble file.
If ncecat is used on files which contain an existing record
dimension, that record dimension is converted to a fixed-length
dimension of the same name and a new record dimension (named
record) is created.
Consider five realizations, 85a.nc, 85b.nc,
... 85e.nc of 1985 predictions from the same climate
model.
Then ncecat 85?.nc 85_ens.nc glues the individual realizations
together into the single file, 85_ens.nc.
If an input variable was dimensioned [lat,lon], it will
have dimensions [record,lat,lon] in the output file.
A restriction of ncecat is that the hyperslabs of the
processed variables must be the same from file to file.
Normally this means all the input files are the same size, and contain
data on different realizations of the same variables.
See ncecat netCDF Ensemble Concatenator, for a complete description
of ncecat.
ncpdq makes it possible to concatenate files along any
dimension, not just the record dimension.
First, use ncpdq to convert the dimension to be concatenated
(i.e., extended with data from other files) into the record dimension.
Second, use ncrcat to concatenate these files.
Finally, if desirable, use ncpdq to revert to the original
dimensionality.
As a concrete example, say that files x_01.nc, x_02.nc,
... x_10.nc contain time-evolving datasets from spatially
adjacent regions.
The time and spatial coordinates are time and x, respectively.
Initially the record dimension is time.
Our goal is to create a single file that contains joins all the
spatially adjacent regions into one single time-evolving dataset.
for idx in 01 02 03 04 05 06 07 08 09 10; do # Bourne Shell
ncpdq -a x,time x_${idx}.nc foo_${idx}.nc # Make x record dimension
done
ncrcat foo_??.nc out.nc # Concatenate along x
ncpdq -a time,x out.nc out.nc # Revert to time as record dimension
Note that ncrcat will not concatenate fixed-length variables, whereas ncecat concatenates both fixed-length and record variables along a new record variable. To conserve system memory, use ncrcat where possible.
The differences between the averagers ncra and ncea are analogous to the differences between the concatenators. ncra is designed for averaging record variables from at least one file, while ncea is designed for averaging fixed length variables from multiple files. ncra performs a simple arithmetic average over the record dimension of all the input files, with each record having an equal weight in the average. ncea performs a simple arithmetic average of all the input files, with each file having an equal weight in the average. Note that ncra cannot average fixed-length variables, but ncea can average both fixed-length and record variables. To conserve system memory, use ncra rather than ncea where possible (e.g., if each input-file is one record long). The file output from ncea will have the same dimensions (meaning dimension names as well as sizes) as the input hyperslabs (see ncea netCDF Ensemble Averager, for a complete description of ncea). The file output from ncra will have the same dimensions as the input hyperslabs except for the record dimension, which will have a size of 1 (see ncra netCDF Record Averager, for a complete description of ncra).
ncflint can interpolate data between or two files. Since no other operators have this ability, the description of interpolation is given fully on the ncflint reference page (see ncflint netCDF File Interpolator). Note that this capability also allows ncflint to linearly rescale any data in a netCDF file, e.g., to convert between differing units.
Occasionally one desires to digest (i.e., concatenate or average)
hundreds or thousands of input files.
Unfortunately, data archives (e.g., NASA EOSDIS) may not
name netCDF files in a format understood by the ‘-n loop’
switch (see Specifying Input Files) that automagically generates
arbitrary numbers of input filenames.
The ‘-n loop’ switch has the virtue of being concise,
and of minimizing the command line.
This helps keeps output file small since the command line is stored
as metadata in the history attribute
(see History Attribute).
However, the ‘-n loop’ switch is useless when there is no
simple, arithmetic pattern to the input filenames (e.g.,
h00001.nc, h00002.nc, ... h90210.nc).
Moreover, filename globbing does not work when the input files are too
numerous or their names are too lengthy (when strung together as a
single argument) to be passed by the calling shell to the NCO
operator
13.
When this occurs, the ANSI C-standard argc-argv
method of passing arguments from the calling shell to a C-program (i.e.,
an NCO operator) breaks down.
There are (at least) three alternative methods of specifying the input
filenames to NCO in environment-limited situations.
The recommended method for sending very large numbers (hundreds or
more, typically) of input filenames to the multi-file operators is
to pass the filenames with the UNIX standard input
feature, aka stdin:
# Pipe large numbers of filenames to stdin
/bin/ls | grep ${CASEID}_'......'.nc | ncecat -o foo.nc
This method avoids all constraints on command line size imposed by
the operating system.
A drawback to this method is that the history attribute
(see History Attribute) does not record the name of any input
files since the names were not passed on the command line.
This makes determining the data provenance at a later date difficult.
To remedy this situation, multi-file operators store the number of
input files in the nco_input_file_number global attribute and the
input file list itself in the nco_input_file_list global attribute
(see File List Attributes).
Although this does not preserve the exact command used to generate the
file, it does retains all the information required to reconstruct the
command and determine the data provenance.
A second option is to use the UNIX xargs command. This simple example selects as input to xargs all the filenames in the current directory that match a given pattern. For illustration, consider a user trying to average millions of files which each have a six character filename. If the shell buffer can not hold the results of the corresponding globbing operator, ??????.nc, then the filename globbing technique will fail. Instead we express the filename pattern as an extended regular expression, ......\.nc (see Subsetting Variables). We use grep to filter the directory listing for this pattern and to pipe the results to xargs which, in turn, passes the matching filenames to an NCO multi-file operator, e.g., ncecat.
# Use xargs to transfer filenames on the command line
/bin/ls | grep ${CASEID}_'......'.nc | xargs -x ncecat -o foo.nc
The single quotes protect the only sensitive parts of the extended
regular expression (the grep argument), and allow shell
interpolation (the ${CASEID} variable substitution) to
proceed unhindered on the rest of the command.
xargs uses the UNIX pipe feature to append the
suitably filtered input file list to the end of the ncecat
command options.
The -o foo.nc switch ensures that the input files supplied by
xargs are not confused with the output file name.
xargs does, unfortunately, have its own limit (usually about
20,000 characters) on the size of command lines it can pass.
Give xargs the ‘-x’ switch to ensure it dies if it
reaches this internal limit.
When this occurs, use either the stdin method above, or the
symbolic link presented next.
Even when its internal limits have not been reached, the xargs technique may not be sophisticated enough to handle all situations. A full scripting language like Perl can handle any level of complexity of filtering input filenames, and any number of filenames. The technique of last resort is to write a script that creates symbolic links between the irregular input filenames and a set of regular, arithmetic filenames that the ‘-n loop’ switch understands. For example, the following Perl script a monotonically enumerated symbolic link to up to one million .nc files in a directory. If there are 999,999 netCDF files present, the links are named 000001.nc to 999999.nc:
# Create enumerated symbolic links
/bin/ls | grep \.nc | perl -e \
'$idx=1;while(<STDIN>){chop;symlink $_,sprintf("%06d.nc",$idx++);}'
ncecat -n 999999,6,1 000001.nc foo.nc
# Remove symbolic links when finished
/bin/rm ??????.nc
The ‘-n loop’ option tells the NCO operator to
automatically generate the filnames of the symbolic links.
This circumvents any OS and shell limits on command line size.
The symbolic links are easily removed once NCO is finished.
One drawback to this method is that the history attribute
(see History Attribute) retains the filename list of the symbolic
links, rather than the data files themselves.
This makes it difficult to determine the data provenance at a later date.
Large datasets are those files that are comparable in size to the amount of random access memory (RAM) in your computer. Many users of NCO work with files larger than 100 MB. Files this large not only push the current edge of storage technology, they present special problems for programs which attempt to access the entire file at once, such as ncea and ncecat. If you work with a 300 MB files on a machine with only 32 MB of memory then you will need large amounts of swap space (virtual memory on disk) and NCO will work slowly, or even fail. There is no easy solution for this. The best strategy is to work on a machine with sufficient amounts of memory and swap space. Since about 2004, many users have begun to produce or analyze files exceeding 2 GB in size. These users should familiarize themselves with NCO's Large File Support (LFS) capabilities (see Large File Support). The next section will increase your familiarity with NCO's memory requirements. With this knowledge you may re-design your data reduction approach to divide the problem into pieces solvable in memory-limited situations.
If your local machine has problems working with large files, try running
NCO from a more powerful machine, such as a network server.
Certain machine architectures, e.g., Cray UNICOS, have special
commands which allow one to increase the amount of interactive memory.
On Cray systems, try to increase the available memory with the
ilimit command.
If you get a memory-related core dump
(e.g., ‘Error exit (core dumped)’) on a GNU/Linux system,
try increasing the process-available memory with ulimit.
The speed of the NCO operators also depends on file size.
When processing large files the operators may appear to hang, or do
nothing, for large periods of time.
In order to see what the operator is actually doing, it is useful to
activate a more verbose output mode.
This is accomplished by supplying a number greater than 0 to the
‘-D debug-level’ (or ‘--debug-level’, or
‘--dbg_lvl’) switch.
When the debug-level is nonzero, the operators report their
current status to the terminal through the stderr facility.
Using ‘-D’ does not slow the operators down.
Choose a debug-level between 1 and 3 for most situations,
e.g., ncea -D 2 85.nc 86.nc 8586.nc.
A full description of how to estimate the actual amount of memory the
multi-file NCO operators consume is given in
Memory Requirements.
Many people use NCO on gargantuan files which dwarf the memory available (free RAM plus swap space) even on today's powerful machines. These users want NCO to consume the least memory possible so that their scripts do not have to tediously cut files into smaller pieces that fit into memory. We commend these greedy users for pushing NCO to its limits!
This section describes the memory NCO requires during operation. The required memory is based on the underlying algorithms. The description below is the memory usage per thread. Users with shared memory machines may use the threaded NCO operators (see OpenMP Threading). The peak and sustained memory usage will scale accordingly, i.e., by the number of threads. Memory consumption patterns of all operators are similar, with the exception of ncap2.
The multi-file operators currently comprise the record operators, ncra and ncrcat, and the ensemble operators, ncea and ncecat. The record operators require much less memory than the ensemble operators. This is because the record operators operate on one single record (i.e., time-slice) at a time, wherease the ensemble operators retrieve the entire variable into memory. Let MS be the peak sustained memory demand of an operator, FT be the memory required to store the entire contents of all the variables to be processed in an input file, FR be the memory required to store the entire contents of a single record of each of the variables to be processed in an input file, VR be the memory required to store a single record of the largest record variable to be processed in an input file, VT be the memory required to store the largest variable to be processed in an input file, VI be the memory required to store the largest variable which is not processed, but is copied from the initial file to the output file. All operators require MI = VI during the initial copying of variables from the first input file to the output file. This is the initial (and transient) memory demand. The sustained memory demand is that memory required by the operators during the processing (i.e., averaging, concatenation) phase which lasts until all the input files have been processed. The operators have the following memory requirements: ncrcat requires MS <= VR. ncecat requires MS <= VT. ncra requires MS = 2FR + VR. ncea requires MS = 2FT + VT. ncbo requires MS <= 3VT (both input variables and the output variable). ncflint requires MS <= 3VT (both input variables and the output variable). ncpdq requires MS <= 2VT (one input variable and the output variable). ncwa requires MS <= 8VT (see below). Note that only variables that are processed, e.g., averaged, concatenated, or differenced, contribute to MS. Variables which do not appear in the output file (see Subsetting Variables) are never read and contribute nothing to the memory requirements.
ncwa consumes between two and seven times the memory of a variable in order to process it. Peak consumption occurs when storing simultaneously in memory one input variable, one tally array, one input weight, one conformed/working weight, one weight tally, one input mask, one conformed/working mask, and one output variable. When invoked, the weighting and masking features contribute up to three-sevenths and two-sevenths of these requirements apiece. If weights and masks are not specified (i.e., no ‘-w’ or ‘-a’ options) then ncwa requirements drop to MS <= 3VT (one input variable, one tally array, and the output variable).
The above memory requirements must be multiplied by the number of threads thr_nbr (see OpenMP Threading). If this causes problems then reduce (with ‘-t thr_nbr’) the number of threads.
ncap2 has unique memory requirements due its ability to process arbitrarily long scripts of any complexity. All scripts acceptable to ncap2 are ultimately processed as a sequence of binary or unary operations. ncap2 requires MS <= 2VT under most conditions. An exception to this is when left hand casting (see Left hand casting) is used to stretch the size of derived variables beyond the size of any input variables. Let VC be the memory required to store the largest variable defined by left hand casting. In this case, MS <= 2VC.
ncap2 scripts are complete dynamic and may be of arbitrary length. A script that contains many thousands of operations, may uncover a slow memory leak even though each single operation consumes little additional memory. Memory leaks are usually identifiable by their memory usage signature. Leaks cause peak memory usage to increase monotonically with time regardless of script complexity. Slow leaks are very difficult to find. Sometimes a malloc() (or new[]) failure is the only noticeable clue to their existance. If you have good reasons to believe that a memory allocation failure is ultimately due to an NCO memory leak (rather than inadequate RAM on your system), then we would be very interested in receiving a detailed bug report.
An overview of NCO capabilities as of about 2006 is in Zender, C. S. (2008), “Analysis of Self-describing Gridded Geoscience Data with netCDF Operators (NCO)”, Environ. Modell. Softw., doi:10.1016/j.envsoft.2008.03.004. This paper is also available at http://dust.ess.uci.edu/ppr/ppr_Zen08_ems.pdf.
NCO performance and scaling for arithmetic operations is described in Zender, C. S., and H. J. Mangalam (2007), “Scaling Properties of Common Statistical Operators for Gridded Datasets”, Int. J. High Perform. Comput. Appl., 21(4), 485-498, doi:10.1177/1094342007083802. This paper is also available at http://dust.ess.uci.edu/ppr/ppr_ZeM07_ijhpca.pdf.
It is helpful to be aware of the aspects of NCO design that can limit its performance:
Many features have been implemented in more than one operator and are described here for brevity. The description of each feature is preceded by a box listing the operators for which the feature is implemented. Command line switches for a given feature are consistent across all operators wherever possible. If no “key switches” are listed for a feature, then that particular feature is automatic and cannot be controlled by the user.
|
Availability: All operators |
|
Availability: ncatted, ncks, ncrename Short options: None Long options: ‘--hdr_pad’, ‘--header_pad’ |
This optimization exploits the netCDF library nc__enddef()
function, which behaves differently with different versions of netCDF.
It will improve speed of future metadata expansion with CLASSIC
and 64bit netCDF files, but not necessarily with NETCDF4
files, i.e., those created by the netCDF interface to the HDF5
library (see Selecting Output File Format).
|
Availability: ncap2, ncbo, ncea, ncecat,
ncflint, ncpdq, ncra, ncrcat,
ncwa Short options: ‘-t’ Long options: ‘--thr_nbr’, ‘--threads’, ‘--omp_num_threads’ |
OMP_NUM_THREADS environment variable, if present, or from the
OS, if not.
NCO may modify thr_nbr according to its own internal
settings before it requests any threads from the system.
Certain operators contain hard-code limits to the number of threads they
request.
We base these limits on our experience and common sense, and to reduce
potentially wasteful system usage by inexperienced users.
For example, ncrcat is extremely I/O-intensive so we restrict
thr_nbr <= 2 for ncrcat.
This is based on the notion that the best performance that can be
expected from an operator which does no arithmetic is to have one thread
reading and one thread writing simultaneously.
In the future (perhaps with netCDF4), we hope to
demonstrate significant threading improvements with operators
like ncrcat by performing multiple simultaneous writes.
Compute-intensive operators (ncap, ncwa and ncpdq)
benefit most from threading.
The greatest increases in throughput due to threading occur on
large datasets where each thread performs millions, at least,
of floating point operations.
Otherwise, the system overhead of setting up threads probably outweighs
the speed enhancements due to SMP parallelism.
However, we have not yet demonstrated that the SMP parallelism
scales well beyond four threads for these operators.
Hence we restrict thr_nbr <= 4 for all operators.
We encourage users to play with these limits (edit file
nco_omp.c) and send us their feedback.
Once the initial thr_nbr has been modified for any operator-specific limits, NCO requests the system to allocate a team of thr_nbr threads for the body of the code. The operating system then decides how many threads to allocate based on this request. Users may keep track of this information by running the operator with dbg_lvl > 0.
By default, threaded operators attach one global attribute,
nco_openmp_thread_number, to any file they create or modify.
This attribute contains the number of threads the operator used to
process the input files.
This information helps to verify that the answers with threaded and
non-threaded operators are equal to within machine precision.
This information is also useful for benchmarking.
|
Availability: All operators |
Extended options, also called long options, are implemented using the system-supplied getopt.h header file, if possible. This provides the getopt_long function to NCO 14.
The syntax of short options (single letter options) is -key value (dash-key-space-value). Here, key is the single letter option name, e.g., ‘-D 2’.
The syntax of long options (multi-letter options) is --long_name value (dash-dash-key-space-value), e.g., ‘--dbg_lvl 2’ or --long_name=value (dash-dash-key-equal-value), e.g., ‘--dbg_lvl=2’. Thus the following are all valid for the ‘-D’ (short version) or ‘--dbg_lvl’ (long version) command line option.
ncks -D 3 in.nc # Short option
ncks --dbg_lvl=3 in.nc # Long option, preferred form
ncks --dbg_lvl 3 in.nc # Long option, alternate form
The last example is preferred for two reasons. First, ‘--dbg_lvl’ is more specific and less ambiguous than ‘-D’. The long option form makes scripts more self documenting and less error prone. Often long options are named after the source code variable whose value they carry. Second, the equals sign = joins the key (i.e., long_name) to the value in an uninterruptible text block. Experience shows that users are less likely to mis-parse commands when restricted to this form.
GNU implements a superset of the POSIX standard which allows any unambiguous truncation of a valid option to be used.
ncks -D 3 in.nc # Short option
ncks --dbg_lvl=3 in.nc # Long option, full form
ncks --dbg=3 in.nc # Long option, unambiguous truncation
ncks --db=3 in.nc # Long option, unambiguous truncation
ncks --d=3 in.nc # Long option, ambiguous truncation
The first four examples are equivalent and will work as expected. The final example will exit with an error since ncks cannot disambiguate whether ‘--d’ is intended as a truncation of ‘--dbg_lvl’, of ‘--dimension’, or of some other long option.
NCO provides many long options for common switches. For example, the debugging level may be set in all operators with any of the switches ‘-D’, ‘--debug-level’, or ‘--dbg_lvl’. This flexibility allows users to choose their favorite mnemonic. For some, it will be ‘--debug’ (an unambiguous truncation of ‘--debug-level’, and other will prefer ‘--dbg’. Interactive users usually prefer the minimal amount of typing, i.e., ‘-D’. We recommend that scripts which are re-usable employ some form of the long options for future maintainability.
This manual generally uses the short option syntax. This is for historical reasons and to conserve space. The remainder of this manual specifies the full long_name of each option. Users are expected to pick the unambiguous truncation of each option name that most suits their taste.
Availability (-n): ncea, ncecat, ncra, ncrcatAvailability ( -p): All operatorsShort options: ‘-n’, ‘-p’ Long options: ‘--nintap’, ‘--pth’, ‘--path’ |
ncra 85.nc 86.nc 87.nc 88.nc 89.nc 8589.nc
ncra 8[56789].nc 8589.nc
ncra -p input-path 85.nc 86.nc 87.nc 88.nc 89.nc 8589.nc
ncra -n 5,2,1 85.nc 8589.nc
The first method (explicitly specifying all filenames) works by brute
force.
The second method relies on the operating system shell to glob
(expand) the regular expression 8[56789].nc.
The shell passes valid filenames which match the expansion to
ncra.
The third method uses the ‘-p input-path’ argument to specify
the directory where all the input files reside.
NCO prepends input-path (e.g.,
/data/usrname/model) to all input-files (but not to
output-file).
Thus, using ‘-p’, the path to any number of input files need only
be specified once.
Note input-path need not end with ‘/’; the ‘/’ is
automatically generated if necessary.
The last method passes (with ‘-n’) syntax concisely describing the entire set of filenames 15. This option is only available with the multi-file operators: ncra, ncrcat, ncea, and ncecat. By definition, multi-file operators are able to process an arbitrary number of input-files. This option is very useful for abbreviating lists of filenames representable as alphanumeric_prefix+numeric_suffix+.+filetype where alphanumeric_prefix is a string of arbitrary length and composition, numeric_suffix is a fixed width field of digits, and filetype is a standard filetype indicator. For example, in the file ccm3_h0001.nc, we have alphanumeric_prefix = ccm3_h, numeric_suffix = 0001, and filetype = nc.
NCO is able to decode lists of such filenames encoded using the
‘-n’ option.
The simpler (3-argument) ‘-n’ usage takes the form
-n file_number,digit_number,numeric_increment
where file_number is the number of files, digit_number is
the fixed number of numeric digits comprising the numeric_suffix,
and numeric_increment is the constant, integer-valued difference
between the numeric_suffix of any two consecutive files.
The value of alphanumeric_prefix is taken from the input file,
which serves as a template for decoding the filenames.
In the example above, the encoding -n 5,2,1 along with the input
file name 85.nc tells NCO to
construct five (5) filenames identical to the template 85.nc
except that the final two (2) digits are a numeric suffix to be
incremented by one (1) for each successive file.
Currently filetype may be either be empty, nc,
cdf, hdf, or hd5.
If present, these filetype suffixes (and the preceding .)
are ignored by NCO as it uses the ‘-n’ arguments to
locate, evaluate, and compute the numeric_suffix component of
filenames.
Recently the ‘-n’ option has been extended to allow convenient
specification of filenames with “circular” characteristics.
This means it is now possible for NCO to automatically
generate filenames which increment regularly until a specified maximum
value, and then wrap back to begin again at a specified minimum value.
The corresponding ‘-n’ usage becomes more complex, taking one or
two additional arguments for a total of four or five, respectively:
-n
file_number,digit_number,numeric_increment[,numeric_max[,numeric_min]]
where numeric_max, if present, is the maximum integer-value of
numeric_suffix and numeric_min, if present, is the minimum
integer-value of numeric_suffix.
Consider, for example, the problem of specifying non-consecutive input
files where the filename suffixes end with the month index.
In climate modeling it is common to create summertime and wintertime
averages which contain the averages of the months June–July–August,
and December–January–February, respectively:
ncra -n 3,2,1 85_06.nc 85_0608.nc
ncra -n 3,2,1,12 85_12.nc 85_1202.nc
ncra -n 3,2,1,12,1 85_12.nc 85_1202.nc
The first example shows that three arguments to the ‘-n’ option
suffice to specify consecutive months (06, 07, 08) which do not
“wrap” back to a minimum value.
The second example shows how to use the optional fourth and fifth
elements of the ‘-n’ option to specify a wrap value to NCO.
The fourth argument to ‘-n’, if present, specifies the maximum
integer value of numeric_suffix.
In this case the maximum value is 12, and will be formatted as
12 in the filename string.
The fifth argument to ‘-n’, if present, specifies the minimum
integer value of numeric_suffix.
The default minimum filename suffix is 1, which is formatted as
01 in this case.
Thus the second and third examples have the same effect, that is, they
automatically generate, in order, the filenames 85_12.nc,
85_01.nc, and 85_02.nc as input to NCO.
|
Availability: All operators Short options: ‘-o’ Long options: ‘--fl_out’, ‘--output’ |
Specifying fl_out with a switch, rather than as a positional argument, allows fl_out to precede input files in the argument list. This is particularly useful with multi-file operators for three reasons. Multi-file operators may be invoked with hundreds (or more) filenames. Visual or automatic location of fl_out in such a list is difficult when the only syntactic distinction between input and output files is their position. Second, specification of a long list of input files may be difficult (see Large Numbers of Files). Making the input file list the final argument to an operator facilitates using xargs for this purpose. Some alternatives to xargs are very ugly and undesirable. Finally, many users are more comfortable specifying output files with ‘-o fl_out’ near the beginning of an argument list. Compilers and linkers are usually invoked this way.
|
Availability: All operators Short options: ‘-p’, ‘-l’ Long options: ‘--pth’, ‘--path’, ‘--lcl’, ‘--local’ |
To access a file via an anonymous FTP server, supply the remote file's URL. FTP is an intrinsically insecure protocol because it transfers passwords in plain text format. Users should access sites using anonymous FTP when possible. Some FTP servers require a login/password combination for a valid user account. NCO allows these transactions so long as the required information is stored in the .netrc file. Usually this information is the remote machine name, login, and password, in plain text, separated by those very keywords, e.g.,
machine dust.ess.uci.edu login zender password bushlied
Eschew using valuable passwords for FTP transactions, since .netrc passwords are potentially exposed to eavesdropping software 16.
SFTP, i.e., secure FTP, uses SSH-based security protocols that solve the security issues associated with plain FTP. NCO supports SFTP protocol access to files specified with a homebrew syntax of the form
sftp://machine.domain.tld:/path/to/filename
Note the second colon following the top-level-domain (tld). This syntax is a hybrid between an FTP URL and a standard remote file syntax.
To access a file using rcp or scp, specify the Internet address of the remote file. Of course in this case you must have rcp or scp privileges which allow transparent (no password entry required) access to the remote machine. This means that ~/.rhosts or ~/ssh/authorized_keys must be set accordingly on both local and remote machines.
To access a file on NCAR's MSS, specify the full
MSS pathname of the remote file.
NCO will attempt to detect whether the local machine has direct
(synchronous) MSS access.
In this case, NCO attempts to use the NCAR
msrcp command
17, or, failing that, /usr/local/bin/msread.
Otherwise NCO attempts to retrieve the MSS file
through the (asynchronous) Masnet Interface Gateway System
(MIGS) using the nrnet command.
The following examples show how one might analyze files stored on remote systems.
ncks -l . ftp://dust.ess.uci.edu/pub/zender/nco/in.nc
ncks -l . sftp://dust.ess.uci.edu:/home/ftp/pub/zender/nco/in.nc
ncks -l . dust.ess.uci.edu:/home/zender/nco/data/in.nc
ncks -l . /ZENDER/nco/in.nc
ncks -l . mss:/ZENDER/nco/in.nc
ncks -l . http://dust.ess.uci.edu/cgi-bin/dods/nph-dods/dodsdata/in.nc
The first example works verbatim if your system is connected to the
Internet and is not behind a firewall.
The second example works if you have sftp access to the
machine dust.ess.uci.edu.
The third example works if you have rcp or scp
access to the machine dust.ess.uci.edu.
The fourth and fifth examples work on NCAR computers with
local access to the msrcp, msread, or
nrnet commands.
The sixth command works if your local version of NCO is
OPeNDAP-enabled (this is fully described in OPeNDAP).
The above commands can be rewritten using the ‘-p input-path’
option as follows:
ncks -p ftp://dust.ess.uci.edu/pub/zender/nco -l . in.nc
ncks -p sftp://dust.ess.uci.edu:/home/ftp/pub/zender/nco -l . in.nc
ncks -p dust.ess.uci.edu:/home/zender/nco -l . in.nc
ncks -p /ZENDER/nco -l . in.nc
ncks -p mss:/ZENDER/nco -l . in.nc
ncks -p http://dust.ess.uci.edu/cgi-bin/dods/nph-dods/dodsdata \
-l . in.nc
Using ‘-p’ is recommended because it clearly separates the input-path from the filename itself, sometimes called the stub. When input-path is not explicitly specified using ‘-p’, NCO internally generates an input-path from the first input filename. The automatically generated input-path is constructed by stripping the input filename of everything following the final ‘/’ character (i.e., removing the stub). The ‘-l output-path’ option tells NCO where to store the remotely retrieved file and the output file. Often the path to a remotely retrieved file is quite different than the path on the local machine where you would like to store the file. If ‘-l’ is not specified then NCO internally generates an output-path by simply setting output-path equal to input-path stripped of any machine names. If ‘-l’ is not specified and the remote file resides on the NCAR MSS system, then the leading character of input-path, ‘/’, is also stripped from output-path. Specifying output-path as ‘-l ./’ tells NCO to store the remotely retrieved file and the output file in the current directory. Note that ‘-l .’ is equivalent to ‘-l ./’ though the latter is recommended as it is syntactically more clear.
The Distributed Oceanographic Data System (DODS) provides useful replacements for common data interface libraries like netCDF. The DODS versions of these libraries implement network transparent access to data via a client-server data access protocol that uses the HTTP protocol for communication. Although DODS-technology originated with oceanography data, it applyies to virtually all scientific data. In recognition of this, the data access protocol underlying DODS (which is what NCO cares about) has been renamed the Open-source Project for a Network Data Access Protocol, OPeNDAP. We use the terms DODS and OPeNDAP interchangeably, and often write OPeNDAP/DODS for now. In the future we will deprecate DODS in favor of DAP or OPeNDAP, as appropriate 18.
NCO may be DAP-enabled by linking
NCO to the OPeNDAP libraries.
This is described in the OPeNDAP documentation and
automagically implemented in NCO build mechanisms
19.
The ./configure mechanism automatically enables NCO as
OPeNDAP clients if it can find the required
OPeNDAP libraries
20.
in the usual locations.
The $DODS_ROOT environment variable may be used to override the
default OPeNDAP library location at NCO
compile-time.
Building NCO with bld/Makefile and the command
make DODS=Y adds the (non-intuitive) commands to link to the
OPeNDAP libraries installed in the $DODS_ROOT
directory.
The file doc/opendap.sh contains a generic script intended to help
users install OPeNDAP before building NCO.
The documentation at the
OPeNDAP Homepage
is voluminous.
Check there and on the
DODS mail lists.
to learn more about the extensive capabilities of OPeNDAP
21.
Once NCO is DAP-enabled the operators are OPeNDAP clients. All OPeNDAP clients have network transparent access to any files controlled by a OPeNDAP server. Simply specify the input file path(s) in URL notation and all NCO operations may be performed on remote files made accessible by a OPeNDAP server. This command tests the basic functionality of OPeNDAP-enabled NCO clients:
% ncks -o ~/foo.nc -C -H -v one -l /tmp \
-p http://dust.ess.uci.edu/cgi-bin/dods/nph-dods/dodsdata in.nc
one = 1
% ncks -H -v one ~/foo.nc
one = 1
The one = 1 outputs confirm (first) that ncks correctly
retrieved data via the OPeNDAP protocol and (second) that
ncks created a valid local copy of the subsetted remote file.
The next command is a more advanced example which demonstrates the real power of OPeNDAP-enabled NCO clients. The ncwa client requests an equatorial hyperslab from remotely stored NCEP reanalyses data of the year 1969. The NOAA OPeNDAP server (hopefully!) serves these data. The local ncwa client then computes and stores (locally) the regional mean surface pressure (in Pa).
ncwa -C -a lat,lon,time -d lon,-10.,10. -d lat,-10.,10. -l /tmp -p \
http://www.cdc.noaa.gov/cgi-bin/nph-nc/Datasets/ncep.reanalysis.dailyavgs/surface \
pres.sfc.1969.nc ~/foo.nc
All with one command! The data in this particular input file also happen to be packed (see Methods and functions), although this is completely transparent to the user since NCO automatically unpacks data before attempting arithmetic.
NCO obtains remote files from the OPeNDAP server (e.g., www.cdc.noaa.gov) rather than the local machine. Input files are first copied to the local machine, then processed. The OPeNDAP server performs data access, hyperslabbing, and transfer to the local machine. This allows the I/O to appear to NCO as if the input files were local. The local machine performs all arithmetic operations. Only the hyperslabbed output data are transferred over the network (to the local machine) for the number-crunching to begin. The advantages of this are obvious if you are examining small parts of large files stored at remote locations.
|
Availability: All operators Short options: ‘-R’ Long options: ‘--rtn’, ‘--retain’ |
Invoking -R disables the default printing behavior of
ncks.
This allows ncks to retrieve remote files without
automatically trying to print them.
See ncks netCDF Kitchen Sink, for more details.
Note that the remote retrieval features of NCO can always be used to retrieve any file, including non-netCDF files, via SSH, anonymous FTP, or msrcp. Often this method is quicker than using a browser, or running an FTP session from a shell window yourself. For example, say you want to obtain a JPEG file from a weather server.
ncks -R -p ftp://weather.edu/pub/pix/jpeg -l . storm.jpg
In this example, ncks automatically performs an anonymous FTP login to the remote machine and retrieves the specified file. When ncks attempts to read the local copy of storm.jpg as a netCDF file, it fails and exits, leaving storm.jpg in the current directory.
If your NCO is DAP-enabled (see OPeNDAP), then you may use NCO to retrieve any files (including netCDF, HDF, etc.) served by an OPeNDAP server to your local machine. For example,
ncks -R -l . -p \
http://www.cdc.noaa.gov/cgi-bin/nph-nc/Datasets/ncep.reanalysis.dailyavgs/surface \
pres.sfc.1969.nc
Note that NCO is never the preffered way to transport files from remote machines. For large jobs, that is best handled by FTP, SSH, or wget. It may occasionally be useful to use NCO to transfer files when your other preferred methods are not available locally.
|
Availability: ncap2, ncbo, ncea,
ncecat, ncflint, ncks, ncpdq,
ncra, ncrcat, ncwa Short options: ‘-3’, ‘-4’ Long options: ‘--3’, ‘--4’, ‘--64bit’, ‘--fl_fmt’, ‘--netcdf4’ |
netCDF supports four types of files: CLASSIC, 64BIT,
NETCDF4, and NETCDF4_CLASSIC,
The CLASSIC format is the traditional 32-bit offset written by
netCDF2 and netCDF3.
As of 2005, most netCDF datasets are in CLASSIC format.
The 64BIT format was added in Fall, 2004.
The NETCDF4 format uses HDF5 as the file storage layer.
The files are (usually) created, accessed, and manipulated using the
traditional netCDF3 API (with numerous extensions).
The NETCDF4_CLASSIC format refers to netCDF4 files created with
the NC_CLASSIC_MODEL mask.
Such files use HDF5 as the back-end storage format (unlike
netCDF3), though they incorporate only netCDF3 features.
Hence NETCDF4_CLASSIC files are perfectly readable by
applications which use only the netCDF3 API and library.
NCO must be built with netCDF4 to write files in the new
NETCDF4 and NETCDF4_CLASSIC formats, and to read files in
the new NETCDF4 format.
Users are advised to use the default CLASSIC format or the
NETCDF4_CLASSIC format until netCDF4 is more widespread.
Widespread support for NETCDF4 format files is not expected for
many years, 2009–2010, say.
If performance or coolness are issues, then use NETCDF4_CLASSIC
instead of CLASSIC format files.
As mentioned above, all operators write use the input file format for
output files unless told otherwise.
Toggling the long option ‘--64bit’ switch (or its
key-value equivalent ‘--fl_fmt=64bit’) produces the
netCDF3 64-bit offset format named 64BIT.
NCO must be built with netCDF 3.6 or higher to produce
a 64BIT file.
Using the ‘-4’ switch (or its long option equivalents
‘--4’ or ‘--netcdf4’), or setting its key-value
equivalent ‘--fl_fmt=netcdf4’ produces a NETCDF4 file
(i.e., HDF).
Casual users are advised to use the default (netCDF3) CLASSIC
format until netCDF 3.6 and netCDF 4.0 are more widespread.
Conversely, operators given the ‘-3’ (or ‘--3’) switch
without arguments will (attempt to) produce netCDF3 CLASSIC
output, even from netCDF4 input files.
These examples demonstrate converting a file from any netCDF format into any other netCDF format (subject to limits of the format):
ncks --fl_fmt=classic in.nc foo_3c.nc # netCDF3 classic
ncks --fl_fmt=64bit in.nc foo_364.nc # netCDF3 64bit
ncks --fl_fmt=netcdf4_classic in.nc foo_4c.nc # netCDF4 classic
ncks --fl_fmt=netcdf4 in.nc foo_4.nc # netCDF4
ncks -3 in.nc foo_3c.nc # netCDF3 classic
ncks --3 in.nc foo_3c.nc # netCDF3 classic
ncks -4 in.nc foo_4.nc # netCDF4
ncks --4 in.nc foo_4.nc # netCDF4
ncks --64 in.nc foo364.nc # netCDF3 64bit
Of course since most operators support these switches, the
“conversions” can be done at the output stage of arithmetic
or metadata processing rather than requiring a separate step.
Producing (netCDF3) CLASSIC or 64BIT files from
NETCDF4_CLASSIC files will always work.
However, producing netCDF3 files from NETCDF4 files will only
work if the output files are not required to contain netCDF4-specific
features.
Note that NETCDF4 and NETCDF4_CLASSIC are the same
binary format.
The latter simply causes a writing application to fail if it attempts to
write a NETCDF4 file that cannot be completely read by the
netCDF3 library.
Conversely, NETCDF4_CLASSIC indicates to a reading application
that all of the file contents are readable with the netCDF3 library.
As of October, 2005, NCO writes no netCDF4-specific data
structures and so always succeeds at writing NETCDF4_CLASSIC
files.
There are at least three ways to discover the format of a netCDF file, i.e., whether it is a classic (32-bit offset) or newer 64-bit offset netCDF3 format, or is netCDF4 format. Each method returns the information using slightly different terminology that becomes easier to understand with practice.
First, examine the end of the first line of global metadata output by ‘ncks -M’:
% ncks -M foo_3c.nc
Opened file foo_3c.nc: dimensions = 19, variables = 261, global atts. = 4,
id = 65536, type = NC_FORMAT_CLASSIC
% ncks -M foo_364.nc
Opened file foo_364.nc: dimensions = 19, variables = 261, global atts. = 4,
id = 65536, type = NC_FORMAT_64BIT
% ncks -M foo_4c.nc
Opened file foo_4c.nc: dimensions = 19, variables = 261, global atts. = 4,
id = 65536, type = NC_FORMAT_NETCDF4_CLASSIC
% ncks -M foo_4.nc
Opened file foo_4.nc: dimensions = 19, variables = 261, global atts. = 4,
id = 65536, type = NC_FORMAT_NETCDF4
This method requires a netCDF4-enabled NCO version 3.9.0+ (i.e., from 2007 or later).
Second, query the file with ‘ncdump -k’:
% ncdump -k foo_3.nc
classic
% ncdump -k foo_364.nc
64-bit-offset
% ncdump -k foo_4c.nc
netCDF-4 classic model
% ncdump -k foo_4.nc
netCDF-4
This method requires a netCDF4-enabled netCDF 3.6.2+ (i.e., from 2007 or later).
The third option uses the POSIX-standard od (octal dump) command:
% od -An -c -N4 foo_3c.nc
C D F 001
% od -An -c -N4 foo_364.nc
C D F 002
% od -An -c -N4 foo_4c.nc
211 H D F
% od -An -c -N4 foo_4.nc
211 H D F
This option works without NCO and ncdump. Values of ‘C D F 001’ and ‘C D F 002’ indicate 32-bit (classic) and 64-bit netCDF3 formats, respectively, while values of ‘211 H D F’ indicate the newer netCDF4 file format.
|
Availability: All operators Short options: none Long options: none |
If you are still interesting in explicit LFS support for netCDF versions prior to 3.6, know that LFS support depends on a complex, interlocking set of operating system 22 and netCDF suppport issues. The netCDF LFS FAQ at http://my.unidata.ucar.edu/content/software/netcdf/faq-lfs.html describes the various file size limitations imposed by different versions of the netCDF standard. NCO and netCDF automatically attempt to configure LFS at build time.
|
Availability: (ncap2), ncbo, ncea,
ncecat, ncflint, ncks, ncpdq,
ncra, ncrcat, ncwa Short options: ‘-v’, ‘-x’ Long options: ‘--variable’, ‘--exclude’ or ‘--xcl’ |
Variables explicitly specified for extraction with ‘-v var[,...]’ must be present in the input file or an error will result. Variables explicitly specified for exclusion with ‘-x -v var[,...]’ need not be present in the input file. Remember, if averaging or concatenating large files stresses your systems memory or disk resources, then the easiest solution is often to use the ‘-v’ option to retain only the most important variables (see Memory Requirements).
Due to its special capabilities, ncap2 interprets the ‘-v’ switch differently (see ncap2 netCDF Arithmetic Processor). For ncap2, the ‘-v’ switch takes no arguments and indicates that only user-defined variables should be output. ncap2 neither accepts nor understands the -x switch.
As of NCO 2.8.1 (August, 2003), variable name arguments of the ‘-v’ switch may contain extended regular expressions. As of NCO 3.9.6 (January, 2009), variable names arguments to ncatted may contain extended regular expressions. For example, ‘-v '^DST'’ selects all variables beginning with the string ‘DST’. Extended regular expressions are defined by the GNU egrep command. The meta-characters used to express pattern matching operations are ‘^$+?.*[]{}|’. If the regular expression pattern matches any part of a variable name then that variable is selected. This capability is called wildcarding, and is very useful for sub-setting large data files.
Because of its wide availability, NCO uses the POSIX
regular expression library regex.
Regular expressions of arbitary complexity may be used.
Since netCDF variable names are relatively simple constructs, only a
few varieties of variable wildcards are likely to be useful.
For convenience, we define the most useful pattern matching operators
here:
Q, Q01–Q99,
Q100, QAA–QZZ, Q_H2O, X_H2O,
Q_CO2, X_CO2.
ncks -v 'Q.?' in.nc # Variables that contain Q
ncks -v '^Q.?' in.nc # Variables that start with Q
ncks -v '^Q+.?.' in.nc # Q, Q0--Q9, Q01--Q99, QAA--QZZ, etc.
ncks -v '^Q..' in.nc # Q01--Q99, QAA--QZZ, etc.
ncks -v '^Q[0-9][0-9]' in.nc # Q01--Q99, Q100
ncks -v '^Q[[:digit:]]{2}' in.nc # Q01--Q99
ncks -v 'H2O$' in.nc # Q_H2O, X_H2O
ncks -v 'H2O$|CO2$' in.nc # Q_H2O, X_H2O, Q_CO2, X_CO2
ncks -v '^Q[0-9][0-9]$' in.nc # Q01--Q99
ncks -v '^Q[0-6][0-9]|7[0-3]' in.nc # Q01--Q73, Q100
ncks -v '(Q[0-6][0-9]|7[0-3])$' in.nc # Q01--Q73
ncks -v '^[a-z]_[a-z]{3}$' in.nc # Q_H2O, X_H2O, Q_CO2, X_CO2
Beware—two of the most frequently used repetition pattern matching operators, ‘*’ and ‘?’, are also valid pattern matching operators for filename expansion (globbing) at the shell-level. Confusingly, their meanings in extended regular expressions and in shell-level filename expansion are significantly different. In an extended regular expression, ‘*’ matches zero or more occurences of the preceding regular expression. Thus ‘Q*’ selects all variables, and ‘Q+.*’ selects all variables containing ‘Q’ (the ‘+’ ensures the preceding item matches at least once). To match zero or one occurence of the preceding regular expression, use ‘?’. Documentation for the UNIX egrep command details the extended regular expressions which NCO supports.
One must be careful to protect any special characters in the regular expression specification from being interpreted (globbed) by the shell. This is accomplish by enclosing special characters within single or double quotes
ncra -v Q?? in.nc out.nc # Error: Shell attempts to glob wildcards
ncra -v '^Q+..' in.nc out.nc # Correct: NCO interprets wildcards
ncra -v '^Q+..' in*.nc out.nc # Correct: NCO interprets, Shell globs
The final example shows that commands may use a combination of variable wildcarding and shell filename expansion (globbing). For globbing, ‘*’ and ‘?’ have nothing to do with the preceding regular expression! In shell-level filename expansion, ‘*’ matches any string, including the null string and ‘?’ matches any single character. Documentation for bash and csh describe the rules of filename expansion (globbing).
|
Availability: ncap2, ncbo, ncea,
ncecat, ncflint, ncks, ncpdq,
ncra, ncrcat, ncwa Short options: ‘-C’, ‘-c’ Long options: ‘--no-coords’, ‘--no-crd’, ‘--crd’, ‘--coords’ |
lat always carry the
values of lat with them into the output-file.
This feature can be disabled with ‘-C’, which causes NCO
to not automatically add coordinates to the variables appearing in the
output-file.
However, using ‘-C’ does not preclude the user from including some
coordinates in the output files simply by explicitly selecting the
coordinates with the -v option.
The ‘-c’ option, on the other hand, is a shorthand way of
automatically specifying that all coordinate variables in the
input-files should appear in the output-file.
Thus ‘-c’ allows the user to select all the coordinate variables
without having to know their names.
Both ‘-c’ and ‘-C’ honor the CF coordinates
convention described in CF Conventions.
|
Availability: ncbo, ncea, ncecat,
ncflint, ncks, ncpdq, ncra,
ncrcat, ncwa Short options: ‘-F’ Long options: ‘--fortran’ |
Consider a file 85.nc containing 12 months of data in the
record dimension time.
The following hyperslab operations produce identical results, a
June-July-August average of the data:
ncra -d time,5,7 85.nc 85_JJA.nc
ncra -F -d time,6,8 85.nc 85_JJA.nc
Printing variable three_dmn_var in file in.nc first with the C indexing convention, then with Fortran indexing convention results in the following output formats:
% ncks -v three_dmn_var in.nc
lat[0]=-90 lev[0]=1000 lon[0]=-180 three_dmn_var[0]=0
...
% ncks -F -v three_dmn_var in.nc
lon(1)=0 lev(1)=100 lat(1)=-90 three_dmn_var(1)=0
...
|
Availability: ncbo, ncea, ncecat,
ncflint, ncks, ncpdq, ncra,
ncrcat, ncwa Short options: ‘-d dim,[min][,[max][,[stride]]]’ Long options: ‘--dimension dim,[min][,[max][,[stride]]]’, ‘--dmn dim,[min][,[max][,[stride]]]’ |
-d dim,[min][,[max][,[stride]]] short
option (or with the same arguments to the ‘--dimension’ or
‘--dmn’ long options).
At least one hyperslab argument (min, max, or stride)
must be present.
The bounds of the hyperslab to be extracted are specified by the
associated min and max values.
A half-open range is specified by omitting either the min or
max parameter.
The separating comma must be present to indicate the omission of one of
these arguments.
The unspecified limit is interpreted as the maximum or minimum value in
the unspecified direction.
A cross-section at a specific coordinate is extracted by specifying only
the min limit and omitting a trailing comma.
Dimensions not mentioned are passed with no reduction in range.
The dimensionality of variables is not reduced (in the case of a
cross-section, the size of the constant dimension will be one).
If values of a coordinate-variable are used to specify a range or
cross-section, then the coordinate variable must be monotonic (values
either increasing or decreasing).
In this case, command-line values need not exactly match coordinate
values for the specified dimension.
Ranges are determined by seeking the first coordinate value to occur in
the closed range [min,max] and including all subsequent
values until one falls outside the range.
The coordinate value for a cross-section is the coordinate-variable
value closest to the specified value and must lie within the range or
coordinate-variable values.
Coordinate values should be specified using real notation with a decimal point required in the value, whereas dimension indices are specified using integer notation without a decimal point. This convention serves only to differentiate coordinate values from dimension indices. It is independent of the type of any netCDF coordinate variables. For a given dimension, the specified limits must both be coordinate values (with decimal points) or dimension indices (no decimal points). The stride option, if any, must be a dimension index, not a coordinate value. See Stride, for more information on the stride option.
User-specified coordinate limits are promoted to double precision values
while searching for the indices which bracket the range.
Thus, hyperslabs on coordinates of type NC_BYTE and
NC_CHAR are computed numerically rather than lexically, so the
results are unpredictable.
The relative magnitude of min and max indicate to the operator whether to expect a wrapped coordinate (see Wrapped Coordinates), such as longitude. If min > max, the NCO expects the coordinate to be wrapped, and a warning message will be printed. When this occurs, NCO selects all values outside the domain [max < min], i.e., all the values exclusive of the values which would have been selected if min and max were swapped. If this seems confusing, test your command on just the coordinate variables with ncks, and then examine the output to ensure NCO selected the hyperslab you expected (coordinate wrapping is currently only supported by ncks).
Because of the way wrapped coordinates are interpreted, it is very
important to make sure you always specify hyperslabs in the
monotonically increasing sense, i.e., min < max
(even if the underlying coordinate variable is monotonically
decreasing).
The only exception to this is when you are indeed specifying a wrapped
coordinate.
The distinction is crucial to understand because the points selected by,
e.g., -d longitude,50.,340., are exactly the complement of the
points selected by -d longitude,340.,50..
Not specifying any hyperslab option is equivalent to specifying full
ranges of all dimensions.
This option may be specified more than once in a single command
(each hyperslabbed dimension requires its own -d option).
|
Availability: ncbo, ncea, ncecat,
ncflint, ncks, ncpdq, ncra,
ncrcat, ncwa Short options: ‘-d dim,[min][,[max][,[stride]]]’ Long options: ‘--dimension dim,[min][,[max][,[stride]]]’, ‘--dmn dim,[min][,[max][,[stride]]]’ |
The stride is specified as the optional fourth argument to the
‘-d’ hyperslab specification:
-d dim,[min][,[max][,[stride]]].
Specify stride as an integer (i.e., no decimal point) following
the third comma in the ‘-d’ argument.
There is no default value for stride.
Thus using ‘-d time,,,2’ is valid but ‘-d time,,,2.0’ and
‘-d time,,,’ are not.
When stride is specified but min is not, there is an
ambiguity as to whether the extracted hyperslab should begin with (using
C-style, 0-based indexes) element 0 or element ‘stride-1’.
NCO must resolve this ambiguity and it chooses element 0
as the first element of the hyperslab when min is not specified.
Thus ‘-d time,,,stride’ is syntactically equivalent to
‘-d time,0,,stride’.
This means, for example, that specifying the operation
‘-d time,,,2’ on the array ‘1,2,3,4,5’ selects the hyperslab
‘1,3,5’.
To obtain the hyperslab ‘2,4’ instead, simply explicitly specify
the starting index as 1, i.e., ‘-d time,1,,2’.
For example, consider a file 8501_8912.nc which contains 60 consecutive months of data. Say you wish to obtain just the March data from this file. Using 0-based subscripts (see C and Fortran Index Conventions) these data are stored in records 2, 14, ... 50 so the desired stride is 12. Without the stride option, the procedure is very awkward. One could use ncks five times and then use ncrcat to concatenate the resulting files together:
for idx in 02 14 26 38 50; do # Bourne Shell
ncks -d time,${idx} 8501_8912.nc foo.${idx}
done
foreach idx (02 14 26 38 50) # C Shell
ncks -d time,${idx} 8501_8912.nc foo.${idx}
end
ncrcat foo.?? 8589_03.nc
rm foo.??
With the stride option, ncks performs this hyperslab extraction in one operation:
ncks -d time,2,,12 8501_8912.nc 8589_03.nc
See ncks netCDF Kitchen Sink, for more information on ncks.
Applying the stride option to the record dimension in ncra and ncrcat makes it possible, for instance, to average or concatenate regular intervals across multi-file input data sets.
ncra -F -d time,3,,12 85.nc 86.nc 87.nc 88.nc 89.nc 8589_03.nc
ncrcat -F -d time,3,,12 85.nc 86.nc 87.nc 88.nc 89.nc 8503_8903.nc
|
Availability: ncbo, ncea, ncecat,
ncflint, ncks, ncpdq, ncra,
ncrcat Short options: ‘-d dim,[min][,[max][,[stride]]]’ Long options: ‘--dimension dim,[min][,[max][,[stride]]]’, ‘--dmn dim,[min][,[max][,[stride]]]’ |
Multislabs overcome some restraints that limit hyperslabs. A single -d option can only specify a contiguous and/or a regularly spaced multi-dimensional data array. Multislabs are constructed from multiple -d options and may therefore have non-regularly spaced arrays. For example, suppose it is desired to operate on all longitudes from 10.0 to 20.0 and from 80.0 to 90.0 degrees. The combined range of longitudes is not selectable in a single hyperslab specfication of the form ‘-d dimension,min,max’ or ‘-d dimension,min,max,stride’ because its elements are irregularly spaced in coordinate space (and presumably in index space too). The multislab specification for obtaining these values is simply the union of the hyperslabs specifications that comprise the multislab, i.e.,
ncks -d lon,10.,20. -d lon,80.,90. in.nc out.nc
ncks -d lon,10.,15. -d lon,15.,20. -d lon,80.,90. in.nc out.nc
Any number of hyperslabs specifications may be chained together to specify the multislab.
Users may specify redundant ranges of indices in a multislab, e.g.,
ncks -d lon,0,4 -d lon,2,9,2 in.nc out.nc
This command retrieves the first five longitudes, and then every other longitude value up to the tenth. Elements 0, 2, and 4 are specified by both hyperslab arguments (hence this is redundant) but will count only once if an arithmetic operation is being performed. This example uses index-based (not coordinate-based) multislabs because the stride option only supports index-based hyper-slabbing. See Stride, for more information on the stride option.
Multislabs are more efficient than the alternative of sequentially performing hyperslab operations and concatenating the results. This is because NCO employs a novel multislab algorithm to minimize the number of I/O operations when retrieving irregularly spaced data from disk. The NCO multislab algorithm retrieves each element from disk once and only once. Thus users may take some shortcuts in specifying multislabs and the algorithm will obtain the intended values. Specifying redundant ranges is not encouraged, but may be useful on occasion and will not result in unintended consequences.
A final example shows the real power of multislabs. Suppose the Q variable contains three dimensional arrays of distinct chemical constituents in no particular order. We are interested in the NOy species in a certain geographic range. Say that NO, NO2, and N2O5 are elements 0, 1, and 5 of the species dimension of Q. The multislab specification might look something like
ncks -d species,0,1 -d species,5 -d lon,0,4 -d lon,2,9,2 in.nc out.nc
Multislabs are powerful because they may be specified for every dimension at the same time. Thus multislabs obsolete the need to execute multiple ncks commands to gather the desired range of data.
|
Availability: ncks Short options: ‘-d dim,[min][,[max][,[stride]]]’ Long options: ‘--dimension dim,[min][,[max][,[stride]]]’, ‘--dmn dim,[min][,[max][,[stride]]]’ |
Assume the domain of the monotonically increasing longitude coordinate
lon is 0 < lon < 360.
ncks will extract a hyperslab which crosses the Greenwich
meridian simply by specifying the westernmost longitude as min and
the easternmost longitude as max.
The following commands extract a hyperslab containing the Saharan desert:
ncks -d lon,340.,50. in.nc out.nc
ncks -d lon,340.,50. -d lat,10.,35. in.nc out.nc
The first example selects data in the same longitude range as the Sahara.
The second example further constrains the data to having the same
latitude as the Sahara.
The coordinate lon in the output-file, out.nc, will
no longer be monotonic!
The values of lon will be, e.g., ‘340, 350, 0, 10, 20, 30,
40, 50’.
This can have serious implications should you run out.nc through
another operation which expects the lon coordinate to be
monotonically increasing.
Fortunately, the chances of this happening are slim, since lon
has already been hyperslabbed, there should be no reason to hyperslab
lon again.
Should you need to hyperslab lon again, be sure to give
dimensional indices as the hyperslab arguments, rather than coordinate
values (see Hyperslabs).
|
Availability: ncbo, ncea, ncecat,
ncflint, ncks, ncpdq, ncra,
ncrcat Short options: ‘-X lon_min,lon_max,lat_min,lat_max’ Long options: ‘--auxiliary lon_min,lon_max,lat_min,lat_max’ |
standard_name attributes, if any, when interpreting
hyperslab and multi-slab options.
Also ‘--auxiliary’.
This switch supports hyperslabbing cell-based grids over coordinate
ranges.
This works on datasets that associate coordinate variables to
grid-mappings using the CF-convention (see CF Conventions)
coordinates and standard_name attributes described
here.
Currently, NCO understands auxiliary coordinate variables
pointed to by the standard_name attributes for latitude and
longitude.
Cells that contain a value within the user-specified range
[lon_min,lon_max,lat_min,lat_max] are
included in the output hyperslab.
A cell-based grid collapses the horizontal spatial information
(latitude and longitude) and stores it along a one-dimensional
coordinate that has a one-to-one mapping to both latitude and longitude
coordinates.
Rectangular (in longitude and latitude) horizontal hyperslabs cannot
be selected using the typical procedure (see Hyperslabs) of
separately specifying ‘-d’ arguments for longitude and latitude.
Instead, when the ‘-X’ is used, NCO learns the names of
the latitude and longitude coordinates by searching the
standard_name attribute of all variables until it finds
the two variables whose standard_name's are “latitude” and
“longitude”, respectively.
This standard_name attribute for latitude and longitude
coordinates follows the CF-convention
(see CF Conventions).
Putting it all together, consider a variable gds_3dvar output from
simulations on a cell-based geodesic grid.
Although the variable contains three dimensions of data (time, latitude,
and longitude), it is stored in the netCDF file with only two dimensions,
time and gds_crd.
% ncks -m -C -v gds_3dvar ~/nco/data/in.nc
gds_3dvar: # dim. = 2, NC_FLOAT, # att. = 4, ID = 38
gds_3dvar dimension 0: time, size = 10 NC_DOUBLE, dim. ID = 18 (CRD)(REC)
gds_3dvar dimension 1: gds_crd, size = 8 NC_FLOAT, dim. ID = 17 (CRD)
gds_3dvar memory size is 10*8*nco_typ_lng(NC_FLOAT) = 80*4 = 320 bytes
gds_3dvar attribute 0: long_name, size = 17 NC_CHAR, value = Geodesic variable
gds_3dvar attribute 1: units, size = 5 NC_CHAR, value = meter
gds_3dvar attribute 2: coordinates, size = 15 NC_CHAR, value = lat_gds lon_gds
The coordinates attribute lists the names of the latitude and
longitude coordinates, lat_gds and lon_gds, respectively.
The coordinates attribute is recommended though optional.
With it, the user can immediately identify which variables contain
the latitude and longitude coordinates.
Without a coordinates attribute it would be unclear at first
glance whether a variable is on a cell-based grid.
In this example, time is a normal record dimension and
gds_crd is the cell-based dimension.
The cell-based grid file must contain two variables whose
standard_name attributes are “latitude”, and “longitude”:
% ncks -m -C -v lat_gds,lon_gds ~/nco/data/in.nc
lat_gds: # dim. = 1, NC_DOUBLE, # att. = 4, ID = 34
lat_gds dimension 0: gds_crd, size = 8 NC_FLOAT, dim. ID = 17 (CRD)
lat_gds memory size is 8*nco_typ_lng(NC_DOUBLE) = 8*8 = 64 bytes
lat_gds attribute 0: long_name, size = 8 NC_CHAR, value = Latitude
lat_gds attribute 1: standard_name, size = 8 NC_CHAR, value = latitude
lat_gds attribute 2: units, size = 6 NC_CHAR, value = degree
lon_gds: # dim. = 1, NC_DOUBLE, # att. = 4, ID = 35
lon_gds dimension 0: gds_crd, size = 8 NC_FLOAT, dim. ID = 17 (CRD)
lon_gds memory size is 8*nco_typ_lng(NC_DOUBLE) = 8*8 = 64 bytes
lon_gds attribute 0: long_name, size = 9 NC_CHAR, value = Longitude
lon_gds attribute 1: standard_name, size = 9 NC_CHAR, value = longitude
lon_gds attribute 2: units, size = 6 NC_CHAR, value = degree
In this example lat_gds and lon_gds represent the
latitude or longitude, respectively, of cell-based variables.
These coordinates (must) have the same single dimension (gds_crd,
in this case) as the cell-based variables.
And the coordinates must be one-dimensional—multidimensional
coordinates will not work.
This infrastructure allows NCO to identify, interpret, and process (e.g., hyperslab) the variables on cell-based grids as easily as it works with regular grids. To time-average all the values between zero and 180 degrees longitude and between plus and minus 30 degress latitude, we use
ncra -O -X 0.,180.,-30.,30. -v gds_3dvar in.nc out.nc
NCO accepts multiple ‘-X’ arguments for cell-based grids multi-slabs, just as it accepts multiple ‘-d’ arguments for multi-slabs of regular coordinates.
ncra -O -X 0.,180.,-30.,30. -X 270.,315.,45.,90. in.nc out.nc
The arguments to ‘-X’ are always interpreted as floating point numbers, i.e., as coordinate values rather than dimension indices so that these two commands produce identical results
ncra -X 0.,180.,-30.,30. in.nc out.nc
ncra -X 0,180,-30,30 in.nc out.nc
In contrast, arguments to ‘-d’ require decimal places to be recognized as coordinates not indices (see Hyperslabs). We recommend always using decimal points with ‘-X’ arguments to avoid confusion.
|
Availability: ncbo, ncea, ncecat,
ncflint, ncks, ncpdq, ncra,
ncrcat, ncwa Short options: ‘-d dim,[min][,[max][,[stride]]]’ Long options: ‘--dimension dim,[min][,[max][,[stride]]]’, ‘--dmn dim,[min][,[max][,[stride]]]’ |
Two examples suffice to demonstrate the power and convenience of UDUnits support. First, consider extraction of a variable containing non-record coordinates with physical dimensions stored in MKS units. In the following example, the user extracts all wavelengths in the visible portion of the spectrum in terms of the units very frequently used in visible spectroscopy, microns:
% ncks -C -H -v wvl -d wvl,"0.4 micron","0.7 micron" in.nc
wvl[0]=5e-07 meter
The hyperslab returns the correct values because the wvl variable
is stored on disk with a length dimension that UDUnits recognizes in the
units attribute.
The automagical algorithm that implements this functionality is worth
describing since understanding it helps one avoid some potential
pitfalls.
First, the user includes the physical units of the hyperslab dimensions
she supplies, separated by a simple space from the numerical values of
the hyperslab limits.
She encloses each coordinate specifications in quotes so that the shell
does not break the value-space-unit string into separate
arguments before passing them to NCO.
Double quotes ("foo") or single quotes ('foo') are equally
valid for this purpose.
Second, NCO recognizes that units translation is requested
because each hyperslab argument contains text characters and non-initial
spaces.
Third, NCO determines whether the wvl is dimensioned
with a coordinate variable that has a units attribute.
In this case, wvl itself is a coordinate variable.
The value of its units attribute is meter.
Thus wvl passes this test so UDUnits conversion is attempted.
If the coordinate associated with the variable does not contain a
units attribute, then NCO aborts.
Fourth, NCO passes the specified and desired dimension strings
(microns are specified by the user, meters are required by
NCO) to the UDUnits library.
Fifth, the UDUnits library that these dimension are commensurate
and it returns the appropriate linear scaling factors to convert from
microns to meters to NCO.
If the units are incommensurate (i.e., not expressible in the same
fundamental MKS units), or are not listed in the UDUnits database, then
NCO aborts since it cannot determine the user's intent.
Finally, NCO uses the scaling information to convert the
user-specified hyperslab limits into the same physical dimensions as
those of the corresponding cooridinate variable on disk.
At this point, NCO can perform a coordinate hyperslab using
the same algorithm as if the user had specified the hyperslab without
requesting units conversion.
The translation and dimensional innterpretation of time coordinates shows a more powerful, and probably more common, UDUnits application. In this example, the user prints all data between the eighth and ninth of December, 1999, from a variable whose time dimension is hours since the year 1900:
% ncks -H -C -v time_udunits -d time_udunits,"1999-12-08 \
12:00:0.0","1999-12-09 00:00:0.0",2 in.nc foo2.nc
time_udunits[1]=876018 hours since 1900-01-01 00:00:0.0
Here, the user invokes the stride (see Stride) capability to obtain every other timeslice. This is possible because the UDUnits feature is additive, not exclusive—it works in conjunction with all other hyperslabbing (see Hyperslabs) options and in all operators which support hyperslabbing. The following example shows how one might average data in a time period spread across multiple input files
ncra -d time,"1939-09-09 12:00:0.0","1945-05-08 00:00:0.0" \
in1.nc in2.nc in3.nc out.nc
Note that there is no excess whitespace before or after the individual
elements of the ‘-d’ argument.
This is important since, as far as the shell knows, ‘-d’ takes
only one command-line argument.
Parsing this argument into its component
dim,[min][,[max][,[stride]]] elements
(see Hyperslabs) is the job of NCO.
When unquoted whitespace is present between these elements, the shell
passes NCO arugment fragments which will not parse as
intended.
NCO implemented support for the UDUnits2 library with version 3.9.2 (August, 2007). The UDUnits2 package supports non-ASCII characters and logarithmic units. We are interested in user-feedback on these features, which are relatively un-tested with NCO.
The UDUnits package documentation describes the supported formats of time dimensions. Among the metadata conventions which adhere to these formats are the Climate and Forecast (CF) Conventions and the Cooperative Ocean/Atmosphere Research Data Service (COARDS) Conventions. The following ‘-d arguments’ extract the same data using commonly encountered time dimension formats:
-d time,"1918-11-11 11:00:0.0","1939-09-09 00:00:0.0"
All of these formats include at least one dash - in a non-leading character position (a dash in a leading character position is a negative sign). NCO assumes that a non-leading dash in a limit string indicates that a UDUnits date conversion is requested.
As of NCO 4.0.0 some of calendar attributes as specified by the CF conventions are supported. The unsupported types default to mixed Gregorian/Julian as defined by UDUunits.
An Example: Consider the following netcdf variable
variables:
double lon_cal(lon_cal) ;
lon_cal:long_name = "lon_cal" ;
lon_cal:units = "days since 1964-2-28 0:0:0" ;
lon_cal:calendar = "365_day" ;
data:
lon_cal = 1,2,3,4,5,6,7,8,9,10;
So the command
"ncks -v lon_cal -d lon_cal,'1964-3-1 0:00:0.0','1964-3-4 00:00:0.0' in.nc out.nc"
Results in the hyperslab lon_cal=1,2,3,4
netCDF variables should always be stored with MKS (i.e., God's) units, so that application programs may assume MKS dimensions apply to all input variables. The UDUnits feature is intended to alleviate some of the NCO user's pain when handling MKS units. It connects users who think in human-friendly units (e.g., miles, millibars, days) to extract data which are always stored in God's units, MKS (e.g., meters, Pascals, seconds). The feature is not intended to encourage writers to store data in esoteric units (e.g., furlongs, pounds per square inch, fortnights).
|
Availability: ncap2, ncbo, ncea,
ncflint, ncpdq, ncra, ncwa Short options: None |
The phrase missing data refers to data points that are missing, invalid, or for any reason not intended to be arithmetically processed in the same fashion as valid data. The NCO arithmetic operators attempt to handle missing data in an intelligent fashion. There are four steps in the NCO treatment of missing data:
NCO follows the convention that missing data should be stored
with the _FillValue specified in the variable's _FillValue
attributes.
The only way NCO recognizes that a variable may
contain missing data is if the variable has a _FillValue
attribute.
In this case, any elements of the variable which are numerically equal
to the _FillValue are treated as missing data.
NCO adopted the behavior that the default attribute name, if
any, assumed to specify the value of data to ignore is _FillValue
with version 3.9.2 (August, 2007).
Prior to that, the missing_value attribute, if any, was assumed to
specify the value of data to ignore.
Supporting both of these attributes simultaneously is not practical.
Hence the behavior NCO once applied to missing_value it now applies
to any _FillValue.
NCO now treats any missing_value as normal data
23.
It has been and remains most advisable to create both _FillValue
and missing_value attributes with identical values in datasets.
Many legacy datasets contain only missing_value attributes.
NCO can help migrating datasets between these conventions.
One may use ncrename (see ncrename netCDF Renamer) to
rename all missing_value attributes to _FillValue:
ncrename -a .missing_value,_FillValue inout.nc
Alternatively, one may use
ncatted (see ncatted netCDF Attribute Editor) to
add a _FillValue attribute to all variables
ncatted -O -a _FillValue,,o,f,1.0e36 inout.nc
Consider a variable var of type var_type with a
_FillValue attribute of type att_type containing the
value _FillValue.
As a guideline, the type of the _FillValue attribute should be
the same as the type of the variable it is attached to.
If var_type equals att_type then NCO
straightforwardly compares each value of var to
_FillValue to determine which elements of var are to be
treated as missing data.
If not, then NCO converts _FillValue from
att_type to var_type by using the implicit conversion rules
of C, or, if att_type is NC_CHAR
24, by typecasting the results of the C function
strtod(_FillValue).
You may use the NCO operator ncatted to change the
_FillValue attribute and all data whose data is
_FillValue to a new value
(see ncatted netCDF Attribute Editor).
When an NCO arithmetic operator processes a variable var
with a _FillValue attribute, it compares each value of
var to _FillValue before performing an operation.
Note the _FillValue comparison imposes a performance penalty
on the operator.
Arithmetic processing of variables which contain the
_FillValue attribute always incurs this penalty, even when
none of the data are missing.
Conversely, arithmetic processing of variables which do not contain the
_FillValue attribute never incurs this penalty.
In other words, do not attach a _FillValue attribute to a
variable which does not contain missing data.
This exhortation can usually be obeyed for model generated data, but it
may be harder to know in advance whether all observational data will be
valid or not.
NCO averagers (ncra, ncea, ncwa) do not count any element with the value _FillValue towards the average. ncbo and ncflint define a _FillValue result when either of the input values is a _FillValue. Sometimes the _FillValue may change from file to file in a multi-file operator, e.g., ncra. NCO is written to account for this (it always compares a variable to the _FillValue assigned to that variable in the current file). Suffice it to say that, in all known cases, NCO does “the right thing”.
It is impossible to determine and store the correct result of a binary operation in a single variable. One such corner case occurs when both operands have differing _FillValue attributes, i.e., attributes with different numerical values. Since the output (result) of the operation can only have one _FillValue, some information may be lost. In this case, NCO always defines the output variable to have the same _FillValue as the first input variable. Prior to performing the arithmetic operation, all values of the second operand equal to the second _FillValue are replaced with the first _FillValue. Then the arithmetic operation proceeds as normal, comparing each element of each operand to a single _FillValue. Comparing each element to two distinct _FillValue's would be much slower and would be no likelier to yield a more satisfactory answer. In practice, judicious choice of _FillValue values prevents any important information from being lost.
|
Availability: ncap2, ncbo, ncea,
ncecat, ncflint, ncks, ncpdq,
ncra, ncrcat, ncwa Short options: none Long options: ‘--cnk_dmn dmn_nm,cnk_sz’, ‘--chunk_dimension dmn_nm,cnk_sz’ , ‘--cnk_map cnk_map’, ‘--chunk_map cnk_map’, ‘--cnk_plc cnk_plc’, ‘--chunk_policy cnk_plc’, ‘--cnk_scl cnk_sz’, ‘--chunk_scalar cnk_sz’ |
All netCDF4-enabled NCO operators that define variables support a plethora of chunksize options. Chunking can significantly accelerate or degrade read/write access to large datasets. Dataset chunking issues are described in detail here.
The NCO chunking implementation is designed to be flexible. Users control three aspects of the chunking implementation. These are known as the chunking policy, chunking map, and chunksize. The first two are high-level mechanisms that apply to an entire file, while the third allows per-dimension specification of parameters. The implementation is a hybrid of the ncpdq packing policies (see ncpdq netCDF Permute Dimensions Quickly), and the hyperslab specifications (see Hyperslabs). Each aspect is intended to have a sensible default, so that most users will only need to set one switch to obtain sensible chunking. Power users can tune the three switches in tandem to obtain optimal performance.
The user specifies the desired chunking policy with the ‘-P’ switch
(or its long option equivalents, ‘--cnk_plc’ and
‘--chunk_policy’) and its cnk_plc argument.
Five chunking policies are currently implemented:
ncchunkncunchunkThe chunking algorithms must know the chunksizes of each dimension of
each variable to be chunked.
The correspondence between the input variable shape and the chunksizes
is called the chunking map.
The user specifies the desired chunking map with the ‘-M’ switch
(or its long option equivalents, ‘--cnk_map’ and
‘--chunk_map’) and its cnk_map argument.
Four chunking maps are currently implemented:
# Debugging
ncks -O -4 -D 4 --cnk_scl=8 ~/nco/data/in.nc ~/foo.nc
ncks -O -4 -D 4 --cnk_scl=8 /data/zender/dstmch90/dstmch90_clm.nc ~/foo.nc
ncks -O -4 -D 4 --cnk_dmn lat,64 --cnk_dmn lon,128 /data/zender/dstmch90/dstmch90_clm.nc ~/foo.nc
ncks -O -4 -D 4 --cnk_plc=uck ~/foo.nc ~/foo.nc
ncks -O -4 -D 4 --cnk_plc=g2d --cnk_map=rd1 --cnk_dmn lat,64 --cnk_dmn lon,128 /data/zender/dstmch90/dstmch90_clm.nc ~/foo.nc
# Chunk data then unchunk it back to its original state:
ncks -O -4 -D 4 --cnk_plc=all ~/nco/data/in.nc ~/foo.nc
ncks -O -4 -D 4 --cnk_plc=uck ~/foo.nc ~/foo.nc
# Final, cleaner examples for manual
ncks --cnk_plc=all in.nc out.nc # Chunk in.nc
ncks --cnk_plc=unchunk in.nc out.nc # Unchunk in.nc
|
Availability: ncap2, ncbo, ncea,
ncecat, ncflint, ncks, ncpdq,
ncra, ncrcat, ncwa Short options: ‘-L’ Long options: ‘--dfl_lvl’, ‘--deflate’ |
All NCO operators that define variables support
the netCDF4 feature of storing variables compressed with Lempel-Ziv
deflation.
The Lempel-Ziv algorithm is a lossless data compression technique.
Activate this deflation with the -L dfl_lvl short option
(or with the same argument to the ‘--dfl_lvl’ or ‘--deflate’
long options).
Specify the deflation level dfl_lvl on a scale from
no deflation (dfl_lvl = 0) to maximum deflation
(dfl_lvl = 9).
Minimal deflation (dfl_lvl = 1) achieves considerable storage
compression with little time penalty.
Higher deflation levels require more time for compression.
File sizes resulting from minimal (dfl_lvl = 1) and maximal
(dfl_lvl = 9) deflation levels typically differ by a few
percent in size.
To compress an entire file using deflation, use
ncks -4 -L 0 in.nc out.nc # No deflation (fast, no time penalty)
ncks -4 -L 1 in.nc out.nc # Minimal deflation (little time penalty)
ncks -4 -L 9 in.nc out.nc # Maximal deflation (much slower)
Unscientific testing shows that deflation compresses typical climate datasets by 30-60%. Packing, a lossy compression technique available for all netCDF files (see Packed data), can easily compress files by 50%. Packed data may be deflated to squeeze datasets by about 80%.
ncks -4 -L 1 in.nc out.nc # Minimal deflation (~30-60% compression)
ncks -4 -L 9 in.nc out.nc # Maximal deflation (~31-63% compression)
ncpdq in.nc out.nc # Standard packing (~50% compression)
ncpdq -4 -L 9 in.nc out.nc # Deflated packing (~80% compression)
ncks prints deflation parameters, if any, to screen (see ncks netCDF Kitchen Sink).
|
Availability: ncap2, ncbo, ncea,
ncflint, ncpdq, ncra, ncwa Short options: None |
The phrase packed data refers to data which are stored in the standard netCDF3 packing format which employs a lossy algorithm. See ncks netCDF Kitchen Sink for a description of deflation, a lossless compression technique available with netCDF4 only. Packed data may be deflated to save additional space.
Packing
The standard netCDF packing algorithm is lossy, and produces data with
the same dynamic range as the original but which requires no more than
half the space to store.
The packed variable is stored (usually) as type NC_SHORT
with the two attributes required to unpack the variable,
scale_factor and add_offset, stored at the original
(unpacked) precision of the variable
25.
Let min and max be the minimum and maximum values
of x.
scale_factor = (max-min)/ndrv
where ndrv is the number of discrete representable values for given type of packed variable. The theoretical maximum value for ndrv is two raised to the number of bits used to store the packed variable. Thus if the variable is packed into type
NC_SHORT, a two-byte
datatype, then there are at most 2^16 = 65536 distinct values
representible.
In practice, the number of discretely representible values is taken
to be one less than the theoretical maximum.
This leaves extra space and solves potential problems with rounding
which can occur during the unpacking of the variable.
Thus for NC_SHORT, ndrv = 65536 - 1 = 65535.
Less often, the variable may be packed into type NC_CHAR,
where ndrv = 256 - 1 = 255, or type NC_INT where
where ndrv = 4294967295 - 1 = 4294967294.
One useful feature of (lossy) netCDF packing algorithm is that
additional, loss-less packing algorithms perform well on top of it.
Unpacking
The unpacking algorithm depends on the presence of two attributes,
scale_factor and add_offset.
If scale_factor is present for a variable, the data are
multiplied by the value scale_factor after the data are read.
If add_offset is present for a variable, then the
add_offset value is added to the data after the data are read.
If both scale_factor and add_offset attributes are
present, the data are first scaled by scale_factor before the
offset add_offset is added.
upk = scale_factor*pck + add_offset = (max-min)*pck/ndrv + 0.5*(min+max)
When
scale_factor and add_offset are used for packing, the
associated variable (containing the packed data) is typically of type
byte or short, whereas the unpacked values are intended to
be of type int, float, or double.
An attribute's scale_factor and add_offset and
_FillValue, if any, should all be of the type intended for the
unpacked data, i.e., int, float or double.
All NCO arithmetic operators understand packed data. The operators automatically unpack any packed variable in the input file which will be arithmetically processed. For example, ncra unpacks all record variables, and ncwa unpacks all variable which contain a dimension to be averaged. These variables are stored unpacked in the output file.
On the other hand, arithmetic operators do not unpack non-processed variables. For example, ncra leaves all non-record variables packed, and ncwa leaves packed all variables lacking an averaged dimension. These variables (called fixed variables) are passed unaltered from the input to the output file. Hence fixed variables which are packed in input files remain packed in output files. Completely packing and unpacking files is easily accomplished with ncpdq (see ncpdq netCDF Permute Dimensions Quickly). Packing and unpacking individual variables may be done with ncpdq and the ncap2 pack() and unpack() functions (see Methods and functions).
|
Availability: ncap2, ncra, ncea, ncwa Short options: ‘-y’ Long options: ‘--operation’, ‘--op_typ’ |
avgsqravgavgsqrmaxminrmsrmssdnsqrtttlThe mathematical definition of each arithmetic operation is given below. See ncwa netCDF Weighted Averager, for additional information on masks and normalization. If an operation type is not specified with ‘-y’ then the operator performs an arithmetic average by default. Averaging is described first so the terminology for the other operations is familiar.
|
Note for HTML users:
|
The definitions of some of these operations are not universally useful. Mostly they were chosen to facilitate standard statistical computations within the NCO framework. We are open to redefining and or adding to the above. If you are interested in having other statistical quantities defined in NCO please contact the NCO project (see Help Requests and Bug Reports).
EXAMPLES
Suppose you wish to examine the variable prs_sfc(time,lat,lon)
which contains a time series of the surface pressure as a function of
latitude and longitude.
Find the minimium value of prs_sfc over all dimensions:
ncwa -y min -v prs_sfc in.nc foo.nc
Find the maximum value of prs_sfc at each time interval for each
latitude:
ncwa -y max -v prs_sfc -a lon in.nc foo.nc
Find the root-mean-square value of the time-series of prs_sfc at
every gridpoint:
ncra -y rms -v prs_sfc in.nc foo.nc
ncwa -y rms -v prs_sfc -a time in.nc foo.nc
The previous two commands give the same answer but ncra is
preferred because it has a smaller memory footprint.
Also, by default, ncra leaves the (degenerate) time
dimension in the output file (which is usually useful) whereas
ncwa removes the time dimension (unless ‘-b’ is
given).
These operations work as expected in multi-file operators.
Suppose that prs_sfc is stored in multiple timesteps per file
across multiple files, say jan.nc, feb.nc,
march.nc.
We can now find the three month maximium surface pressure at every point.
ncea -y max -v prs_sfc jan.nc feb.nc march.nc out.nc
It is possible to use a combination of these operations to compute the variance and standard deviation of a field stored in a single file or across multiple files. The procedure to compute the temporal standard deviation of the surface pressure at all points in a single file in.nc involves three steps.
ncwa -O -v prs_sfc -a time in.nc out.nc
ncbo -O -v prs_sfc in.nc out.nc out.nc
ncra -O -y rmssdn out.nc out.nc
First construct the temporal mean of prs_sfc in the file
out.nc.
Next overwrite out.nc with the anomaly (deviation from the mean).
Finally overwrite out.nc with the root-mean-square of itself.
Note the use of ‘-y rmssdn’ (rather than ‘-y rms’) in the
final step.
This ensures the standard deviation is correctly normalized by one fewer
than the number of time samples.
The procedure to compute the variance is identical except for the use of
‘-y var’ instead of ‘-y rmssdn’ in the final step.
ncap2 can also compute statistics like standard deviations. Brute-force implementation of formulae is one option, e.g.,
ncap2 -s 'prs_sfc_sdn=sqrt((prs_sfc-prs_sfc.avg($time)^2).total($time))/($time.size-1)'
in.nc out.nc
The operation may, of course, be broken into multiple steps in order to archive intermediate quantities, such as the time-anomalies
ncap2 -s 'prs_sfc_anm=prs_sfc-prs_sfc.avg($time)' \
-s 'prs_sfc_sdn=sqrt((prs_sfc_anm^2).total($time))/($time.size-1)' \
in.nc out.nc
ncap2 supports intrinsic standard deviation functions (see Operation Types) which simplify the above expression to
ncap2 -s 'prs_sfc_sdn=(prs_sfc-prs_sfc.avg($time)).rmssdn($time)' in.nc out.nc
These instrinsic functions compute the answer quickly and concisely.
The procedure to compute the spatial standard deviation of a field in a single file in.nc involves three steps.
ncwa -O -v prs_sfc,gw -a lat,lon -w gw in.nc out.nc
ncbo -O -v prs_sfc,gw in.nc out.nc out.nc
ncwa -O -y rmssdn -v prs_sfc -a lat,lon -w gw out.nc out.nc
First the appropriately weighted (with ‘-w gw’) spatial mean values are written to the output file. This example includes the use of a weighted variable specified with ‘-w gw’. When using weights to compute standard deviations one must remember to include the weights in the initial output files so that they may be used again in the final step. The initial output file is then overwritten with the gridpoint deviations from the spatial mean. Finally the root-mean-square of the appropriately weighted spatial deviations is taken.
The ncap2 solution to the spatially-weighted standard deviation problem is
ncap2 -s 'prs_sfc_sdn=(prs_sfc*gw-prs_sfc*gw.avg($lat,$lon)).rmssdn($lat,$lon)' \
in.nc out.nc
Be sure to multiply the variable by the weight prior to computing the the anomalies and the standard deviation.
The procedure to compute the standard deviation of a time-series across multiple files involves one extra step since all the input must first be collected into one file.
ncrcat -O -v tpt in.nc in.nc foo1.nc
ncwa -O -a time foo1.nc foo2.nc
ncbo -O -v tpt foo1.nc foo2.nc foo2.nc
ncra -O -y rmssdn foo2.nc out.nc
The first step assembles all the data into a single file. This may require a lot of temporary disk space, but is more or less required by the ncbo operation in the third step.
|
Availability: ncap2, ncbo, ncea,
ncra, ncwa Short options: None |
NC_SHORT (two bytes) to NC_DOUBLE (eight bytes).
Type conversion is automatic when the language carries out this
promotion according to an internal set of rules without explicit user
intervention.
In contrast, manual type conversion refers to explicit user commands to
change the type of a variable or attribute.
Most type conversion happens automatically, yet there are situations in
which manual type conversion is advantageous.
As a general rule, automatic type conversions should be avoided for at
least two reasons.
First, type conversions are expensive since they require creating
(temporary) buffers and casting each element of a variable from
the type it was stored at to some other type.
Second, the dataset's creator probably had a good reason
for storing data as, say, NC_FLOAT rather than NC_DOUBLE.
In a scientific framework there is no reason to store data with more
precision than the observations were made.
Thus NCO tries to avoid performing automatic type conversions
when performing arithmetic.
Automatic type conversion during arithmetic in the languages C and
Fortran is performed only when necessary.
All operands in an operation are converted to the most precise type
before the operation takes place.
However, following this parsimonious conversion rule dogmatically
results in numerous headaches.
For example, the average of the two NC_SHORTs 17000s and
17000s results in garbage since the intermediate value which
holds their sum is also of type NC_SHORT and thus cannot
represent values greater than 32,767
26.
There are valid reasons for expecting this operation to succeed and
the NCO philosophy is to make operators do what you want, not
what is most pure.
Thus, unlike C and Fortran, but like many other higher level interpreted
languages, NCO arithmetic operators will perform automatic type
conversion when all the following conditions are met
27:
NC_BYTE, NC_CHAR,
NC_SHORT, or NC_INT.
Type NC_DOUBLE is not type converted because there is no type of
higher precision to convert to.
Type NC_FLOAT is not type converted because, in our judgement,
the performance penalty of always doing so would outweigh the (extremely
rare) potential benefits.
When these criteria are all met, the operator promotes the variable in
question to type NC_DOUBLE, performs all the arithmetic
operations, casts the NC_DOUBLE type back to the original type,
and finally writes the result to disk.
The result written to disk may not be what you expect, because of
incommensurate ranges represented by different types, and because of
(lack of) rounding.
First, continuing the above example, the average (e.g., ‘-y avg’)
of 17000s and 17000s is written to disk as 17000s.
The type conversion feature of NCO makes this possible since
the arithmetic and intermediate values are stored as NC_DOUBLEs,
i.e., 34000.0d and only the final result must be represented
as an NC_SHORT.
Without the type conversion feature of NCO, the average would
have been garbage (albeit predictable garbage near -15768s).
Similarly, the total (e.g., ‘-y ttl’) of 17000s and
17000s written to disk is garbage (actually -31536s) since
the final result (the true total) of 34000 is outside the range
of type NC_SHORT.
Type conversions use the floor function to convert floating point
number to integers.
Type conversions do not attempt to round floating point numbers to the
nearest integer.
Thus the average of 1s and 2s is computed in double
precisions arithmetic as
(1.0d + 1.5d)/2) = 1.5d.
This result is converted to NC_SHORT and stored on disk as
floor(1.5d) = 1s
28.
Thus no "rounding up" is performed.
The type conversion rules of C can be stated as follows:
If n is an integer then any floating point value x
satisfying
n <= x < n+1
will have the value n when converted to an integer.
ncap2 provides intrinsic functions for performing manual type
conversions.
This, for example, converts variable tpt to external type
NC_SHORT (a C-type short), and variable prs to
external type NC_DOUBLE (a C-type double).
ncap2 -s 'tpt=short(tpt);prs=double(prs)' in.nc out.nc
See ncap2 netCDF Arithmetic Processor, for more details.
|
Availability: All operators Short options: ‘-O’, ‘-A’ Long options: ‘--ovr’, ‘--overwrite’, ‘--apn’, ‘--append’ |
|
Availability: All operators Short options: ‘-h’ Long options: ‘--hst’, ‘--history’ |
history global attribute to
any file they create or modify.
The history attribute consists of a timestamp and the full string
of the invocation command to the operator, e.g., ‘Mon May 26 20:10:24
1997: ncks in.nc foo.nc’.
The full contents of an existing history attribute are copied
from the first input-file to the output-file.
The timestamps appear in reverse chronological order, with the most
recent timestamp appearing first in the history attribute.
Since NCO and many other netCDF operators adhere to the
history convention, the entire data processing path of a given
netCDF file may often be deduced from examination of its history
attribute.
As of May, 2002, NCO is case-insensitive to the spelling
of the history attribute name.
Thus attributes named History or HISTORY (which are
non-standard and not recommended) will be treated as valid history
attributes.
When more than one global attribute fits the case-insensitive search
for "history", the first one found will be used.
history attribute
To avoid information overkill, all operators have an optional switch
(‘-h’, ‘--hst’, or ‘--history’) to override
automatically appending the history attribute
(see ncatted netCDF Attribute Editor).
Note that the ‘-h’ switch also turns off writing the
nco_input_file_list attribute for multi-file operators
(see File List Attributes).
|
Availability: ncea, ncecat, ncra, ncrcat Short options: ‘-H’ Long options: ‘--fl_lst_in’, ‘--file_list’ |
history
attribute no longer contains the exact command by which the file
was created.
NCO solves this dilemma by archiving input file list
attributes.
When the input file list to a multi-file operator is specified
via stdin, the operator, by default, attaches two global
attributes to any file they create or modify.
The nco_input_file_number global attribute contains the number of
input files, and nco_input_file_list contains the file names,
specified as standard input to the multi-file operator.
This information helps to verify that all input files the user thinks
were piped through stdin actually arrived.
Without the nco_input_file_list attribute, the information is lost
forever and the “chain of evidence” would be broken.
The ‘-H’ switch overrides (turns off) the default behavior of
writing the input file list global attributes when input is from
stdin.
The ‘-h’ switch does this too, and turns off the history
attribute as well (see History Attribute).
Hence both switches allows space-conscious users to avoid storing what
may amount to many thousands of filenames in a metadata attribute.
|
Availability: ncbo, ncea, ncecat,
ncflint, ncra, ncwa Short options: None |
Conventions attribute (e.g., ‘CF-1.0’).
We refer to all such data collectively as CF data.
Skip this section if you never work with CF data.
The CF netCDF conventions are described at http://www.cgd.ucar.edu/cms/eaton/cf-metadata/CF-1.0.html. Most CF netCDF conventions are transparent to NCO 29. There are no known pitfalls associated with using any NCO operator on files adhering to these conventions 30. However, to facilitate maximum user friendliness, NCO does treat certain variables in some CF files specially. The special functions are not required by the CF netCDF conventions, but experience shows they simplify data analysis.
Currently, NCO determines whether a datafile is a
CF output datafile simply by checking whether value of the
global attribute Conventions (if it exists) equals
‘CF-1.0’ or ‘NCAR-CSM’.
Should Conventions equal either of these in the (first)
input-file, NCO will attempt to treat certain variables
specially, because of their meaning in CF files.
NCO will not average the following variables often found in
CF files:
ntrm, ntrn, ntrk, ndbase, nsbase,
nbdate, nbsec, mdt, mhisf.
These variables contain scalar metadata such as the resolution of the
host geophysical model and it makes no sense to change their values.
Furthermore, the ncbo operator does not operate on (i.e., add,
subtract, etc.) the following variables:
ORO,
area,
datesec,
date,
gw,
hyai,
hyam,
hybi.
hybm,
lat_bnds,
lon_bnds,
msk_*.
These variables represent the Gaussian weights, the orography field,
time fields, hybrid pressure coefficients, and latititude/longitude
boundaries.
We call these fields non-coordinate grid properties.
Coordinate grid properties are easy to identify because they are
coordinate variables such as latitude and longitude.
Users usually want all grid properties to remain unaltered in the
output file.
To be treated as a grid property, the variable name must exactly
match a name in the above list, or be a coordinate variable.
The handling of msk_* is exceptional in that any variable
name beginning with the string msk_ is considered to be a
“mask” and is thus preserved (not operated on arithmetically).
You must spoof NCO if you would like any grid properties
or other special CF fields processed normally.
For example rename the variables first with ncrename,
or alter the Conventions attribute.
NCO supports the CF coordinates convention described
here.
This convention allows variables to specify additional coordinates
(including multidimensional coordinates) in a space-separated string
attribute named coordinates.
NCO attaches any such coordinates to the extraction list along with
variable and its usual (one-dimensional) coordinates, if any.
These auxiliary coordinates are subject to the user-specified overrides
described in Subsetting Coordinate Variables.
|
Availability: ncrcat Short options: None |
base_time, and a record variable, time_offset.
Subtle but serious problems can arise when these type of files are
just blindly concatenated.
Therefore ncrcat has been specially programmed to be able to
chain together consecutive ARM input-files and produce
and an output-file which contains the correct time information.
Currently, ncrcat determines whether a datafile is an
ARM datafile simply by testing for the existence of the
variables base_time, time_offset, and the dimension
time.
If these are found in the input-file then ncrcat will
automatically perform two non-standard, but hopefully useful,
procedures.
First, ncrcat will ensure that values of time_offset
appearing in the output-file are relative to the base_time
appearing in the first input-file (and presumably, though not
necessarily, also appearing in the output-file).
Second, if a coordinate variable named time is not found in the
input-files, then ncrcat automatically creates the
time coordinate in the output-file.
The values of time are defined by the ARM conventions
time = base_time + time_offset.
Thus, if output-file contains the time_offset
variable, it will also contain the time coordinate.
A short message is added to the history global attribute
whenever these ARM-specific procedures are executed.
|
Availability: All operators Short options: ‘-r’ Long options: ‘--revision’, ‘--version’, or ‘--vrs’ |
3.9.5.
Using ‘-r’ on, say, ncks, produces something like
‘NCO netCDF Operators version "3.9.5" last modified 2008/05/11 built May 12 2008 on neige by zender
Copyright (C) 1995--2008 Charlie Zender
ncks version 20090918’.
This tells you that ncks contains all patches up to version
3.9.5, which dates from May 11, 2008.
This chapter presents reference pages for each of the operators individually. The operators are presented in alphabetical order. All valid command line switches are included in the syntax statement. Recall that descriptions of many of these command line switches are provided only in Common features, to avoid redundancy. Only options specific to, or most useful with, a particular operator are described in any detail in the sections below.
| ncap2 understands a relatively full-featured language of operations, including loops, conditionals, arrays, and math functions. ncap2 is the most rapidly changing NCO operator and its documentation is incomplete. The distribution file data/ncap2_tst.nco contains an up-to-date overview of its syntax and capabilities. The data/*.nco distribution files (especially bin_cnt.nco, psd_wrf.nco, and rgr.nco) contain in-depth examples of ncap2 solutions to complex problems. |
SYNTAX
ncap2 [-3] [-4] [-6] [-A] [-C] [-c] [-D dbg] [-F] [-f] [-L dfl_lvl]
[-l path] [-O] [-o output-file] [-p path] [-R] [-r]
[-s algebra] [-S fl.nco] [-t thr_nbr] [-v]
input-file [output-file]
DESCRIPTION
ncap2 arithmetically processes netCDF files 31. The processing instructions are contained either in the NCO script file fl.nco or in a sequence of command line arguments. The options ‘-s’ (or long options ‘--spt’ or ‘--script’) are used for in-line scripts and ‘-S’ (or long options ‘--fl_spt’ or ‘--script-file’) are used to provide the filename where (usually multiple) scripting commands are pre-stored. ncap2 was written to perform arbitrary algebraic transformations of data and archive the results as easily as possible. See Missing Values, for treatment of missing values. The results of the algebraic manipulations are called derived fields.
Unlike the other operators, ncap2 does not accept a list of variables to be operated on as an argument to ‘-v’ (see Subsetting Variables). Rather, the ‘-v’ switch takes no arguments and indicates that ncap2 should output only user-defined variables. ncap2 neither accepts nor understands the -x switch.
Defining new variables in terms of existing variables is a powerful feature of ncap2. Derived fields inherit the metadata (i.e., attributes) of their ancestors, if any, in the script or input file. When the derived field is completely new (no identically-named ancestors exist), then it inherits the metadata (if any) of the left-most variable on the right hand side of the defining expression. This metadata inheritance is called attribute propagation. Attribute propagation is intended to facilitate well-documented data analysis, and we welcome suggestions to improve this feature.
Mastering ncap2 is relatively simple. Each valid statement statement consists of standard forward algebraic expression. The fl.nco, if present, is simply a list of such statements, whitespace, and comments. The syntax of statements is most like the computer language C. The following characteristics of C are preserved:
[] characters;
/* */ characters.
Single line comments are preceded by // characters.
#include script.
Note that the #include command is not followed by a semi-colon
because it is a pre-processor directive, not an assignment statement.
The filename script is interpreted relative to the run directory.
@ is used to delineate an attribute name from a
variable name.
Expressions are the fundamental building block of ncap2. Expressions are composed of variables, numbers, literals, and attributes. The following C operators are “overloaded” and work with scalars and multi-dimensional arrays:
Arithmetic Operators: * / % + - ^
Binary Operators: > >= < <= == != == || && >> <<
Unary Operators: + - ++ -- !
Conditional Operator: exp1 ? exp2 : exp3
Assign Operators: = += -= /= *=
In the following section a variable also refers to a number literal which is read in as a scalar variable:
Arithmetic and Binary Operators
Consider var1 'op' var2
Precision
NC_FLOAT, the result is NC_FLOAT.
When either type is NC_DOUBLE, the result is also NC_DOUBLE.
Rank
Even though the logical operators return True(1) or False(0) they are treated in the same way as the arithmetic operators with regard to precision and rank.
examples:
dimensions: time=10, lat=2, lon=4
Suppose we have the two variables:
double P(time,lat,lon);
float PZ0(lon,lat); // PZ0=1,2,3,4,5,6,7,8;
Consider now the expression:
PZ=P-PZ0
PZ0 is made to conform to P and the result is
PZ0 =
1,3,5,7,2,4,6,8,
1,3,5,7,2,4,6,8,
1,3,5,7,2,4,6,8,
1,3,5,7,2,4,6,8,
1,3,5,7,2,4,6,8,
1,3,5,7,2,4,6,8,
1,3,5,7,2,4,6,8,
1,3,5,7,2,4,6,8,
1,3,5,7,2,4,6,8,
1,3,5,7,2,4,6,8,
Once the expression is evaluated then PZ will be of type double;
Consider now
start=four-att_var@double_att; // start =-69 and is of type intger;
four_pow=four^3.0f // four_pow=64 and is of type float
three_nw=three_dmn_var_sht*1.0f; // type is now float
start@n1=att_var@short_att*att_var@int_att;
// start@n1=5329 and is type int
Binary Operators
Unlike C the binary operators return an array of values. There is no such thing as short circuiting with the AND/OR operators. Missing values are carried into the result in the same way they are with the arithmetic operators. When an expression is evaluated in an if() the missing values are treated as true.
The Binary operators are,in order of precedence:
! Logical Not
----------------------------
<< Less Than Selection
>> Greater Than Selection
----------------------------
> Greater than
>= Greater than or equal to
< Less than
<= Less than or equal to
----------------------------
== Equal to
!= Not equal to
----------------------------
&& Logical AND
----------------------------
|| Logical OR
----------------------------
To see all operators: see Operators precedence and associativity
examples:
tm1= time>2 && time <7; // tm1 = 0, 0, 1, 1, 1, 1, 0, 0, 0, 0 ; type double;
tm2= time==3 || time>=6; // tm2 = 0, 0, 1, 0, 0, 1, 1, 1, 1, 1 ; type double
tm3= int(!tm1); // tm3= 1, 1, 0, 0, 0, 0, 1, 1, 1, 1 ; type int
tm4= tm1 && tm2; // tm4= 0, 0, 1, 0, 0, 1, 0, 0, 0, 0 ; type double;
tm5= !tm4; // tm5= 1, 1, 0, 1, 1, 0, 1, 1, 1, 1 ; type double;
Regular Assign Operator
var1 '=' exp1
If var1 doesn't already exist in Output then var1 is written to Output with the values and dimensions from expr1. If var1 already exists in Output, then the only requirement on expr1 is that the number of elements must match the number already on disk. The type of expr1 is converted if necessary to the disk type.
Other Assign Operators +=,-=,*=./=
var1 'ass_op' exp1
if exp1 is a variable and it doesn't conform to var1 then an attempt is made to make it conform to var1. If exp1 is an attribute it must have unity size or else have the same number of elements as var1. If expr1 has a different type to var1 the it is converted to the var1 type.
example:
z1=four+=one*=10 // z1=14 four=14 one=10;
time-=2 // time= -1,0,1,2,3,4,5,6,7,8
Increment/ Decrement Operators
These work in a similar fashion to their regular C counterparts. If say the variable "four" is input only then the statement "++four" effectively means -read four from input increment each element by one , then write the new values to Output;
example:
n2=++four; n2=5, four=5
n3=one--+20; n3=21 one=0;
n4=--time; n4=time=0.,1.,2.,3.,4.,5.,6.,7.,8.,9.;
Conditional Operator ?:
exp1 ? exp2 : exp3
The conditional operator ( or ternary Operator) is nice and succinct way of writing an if/then/else. If exp1 evaluates to true then exp2 is returned else exp3 is returned.
example
weight_avg= weight.avg();
weight_avg@units= (weight_avg ==1 ? "kilo" : "kilos");
PS_nw= PS - (PS.min() >100000 ? 100000 : 0 );
example:
RDM2= RDM >>100.0; RDM2=100,100,100,100,126,126,100,100,100, 100 ; // type double
RDM2= RDM <<90s; RDM3=1, 9, 36, 84, 90, 90, 84, 36, 9, 1 ; // type int
Dimensions can be defined in Output using the defdim() function.
defdim("cnt",10);
This dimension can then be subsequently referred to in method arguments and a left hand cast by prefixing it with a dollar e.g
new_var[$cnt]=time;
temperature[$time,$lat,$lon]=35.5;
temp_avg=temperature.avg($time);
To refer to the dimension size in an expression use the size method.
time_avg=time.total() / $time.size;
Increase size of new var by one and set new member to zero;
defdim("cnt_grw", $cnt.size+1);
new_var_grw[$cnt_grw]=0.0;
new_var_grw( 0:($cnt_grw.size-2))=new_var;
Dimension Abbreviations
Its possible to use dimension abbreviations as method arguments.
$0 is the first dimension of the variable
$1 is the second dimension of the variable
$n is the n+1 dimension of the variable
consider the variables:
float four_dmn_rec_var(time,lat,lev,lon);
double three_dmn_var_dbl(time,lat,lon);
four_nw=four_dmn_rev_var.reverse($time,$lon)
four_nw=four_dmn_rec_var.reverse($0,$3);
four_avg=four_dmn_rec_var.avg($lat,$lev);
four_avg=four_dmn_rec_var.avg($1,$2);
three_mw=three_dmn_var_dbl.permute($time,$lon,$lat);
three_mw=three_dmn_var_dbl.permute($0,$2,$1);
ID Quoting
If the dim name contains non-regular characters use ID quoting. See see ID Quoting
defdim("a--list.A",10);
A1['$a--list.A']=30.0;
GOTCHA
It is not possible to manually define in Output any dimensions that exist in Input. When a variable from Input appears in an expression or statement its dimensions in Input are automagically copied to Output (if they are not already present)
The following examples demonstrate the utility of the left hand casting ability of ncap2. Consider first this simple, artificial, example. If lat and lon are one dimensional coordinates of dimensions lat and lon, respectively, then addition of these two one-dimensional arrays is intrinsically ill-defined because whether lat_lon should be dimensioned lat by lon or lon by lat is ambiguous (assuming that addition is to remain a commutative procedure, i.e., one that does not depend on the order of its arguments). Differing dimensions are said to be orthogonal to one another, and sets of dimensions which are mutually exclusive are orthogonal as a set and any arithmetic operation between variables in orthogonal dimensional spaces is ambiguous without further information.
The ambiguity may be resolved by enumerating the desired dimension ordering of the output expression inside square brackets on the left hand side (LHS) of the equals sign. This is called left hand casting because the user resolves the dimensional ordering of the RHS of the expression by specifying the desired ordering on the LHS.
ncap2 -s 'lat_lon[lat,lon]=lat+lon' in.nc out.nc
ncap2 -s 'lon_lat[lon,lat]=lat+lon' in.nc out.nc
The explicit list of dimensions on the LHS, [lat,lon]
resolves the otherwise ambiguous ordering of dimensions in
lat_lon.
In effect, the LHS casts its rank properties onto the
RHS.
Without LHS casting, the dimensional ordering of lat_lon
would be undefined and, hopefully, ncap2 would print an error
message.
Consider now a slightly more complex example.
In geophysical models, a coordinate system based on
a blend of terrain-following and density-following surfaces is
called a hybrid coordinate system.
In this coordinate system, four variables must be manipulated to
obtain the pressure of the vertical coordinate:
PO is the domain-mean surface pressure offset (a scalar),
PS is the local (time-varying) surface pressure (usually two
horizontal spatial dimensions, i.e. latitude by longitude), hyam
is the weight given to surfaces of constant density (one spatial
dimension, pressure, which is orthogonal to the horizontal
dimensions), and hybm is the weight given to surfaces of
constant elevation (also one spatial dimension).
This command constructs a four-dimensional pressure prs_mdp
from the four input variables of mixed rank and orthogonality:
ncap2 -s 'prs_mdp[time,lat,lon,lev]=P0*hyam+PS*hybm' in.nc out.nc
Manipulating the four fields which define the pressure in a hybrid coordinate system is easy with left hand casting.
Hyperslabs in ncap2 are a bit more limited than hyperslabs with the other nco operators. There is no per-se multi-slabs, wrapped co-ordinates, negative stride or co-ordinate value limits. However with a bit of syntactic magic they are all are possible.
var1( hyper_arg1, hyper_arg2 .. hyper_argN)
A hyperslab argument is specified using the following notation
start:end:stride
if "start" is omitted - then default = 0
if "end" is omitted - default = dimension size less one
if "stride" is omitted - default = 1
If a single value is present then it is assumed that that dimension collapses to a single value (ie a cross-section). The number of hyperslab arguments MUST be equal to the number of dimensions of the variable.
Hyperslabs on the Right Hand Side of an assign
A simple 1D example:
($time.size=10)
od[$time]={20,22,24,26,28,30,32,34,36,38};
od(7); // 34
od(7:); // 34,36,38
od(:7); // 20, 22, 24, 26, 28, 30, 32, 34
od(::4); // 20.28,36
od(1:6:2) // 22,26,30
od(:) // 20,22,24,26,28,30,32,34,36,38
A more complex 3D example
($lat.size=2, $lon.size=4 )
th[$time,$lat,$lon]=
{1, 2, 3, 4, 5, 6, 7, 8,
9,10,11,12,13,14,15,16,
17,18,19,20,21,22,23,24,
-99,-99,-99,-99,-99,-99,-99,-99,
33,34,35,36,37,38,39,40,
41,42,43,44,45,46,47,48,
49,50,51,52,53,54,55,56,
-99,58,59,60,61,62,63,64,
65,66,67,68,69,70,71,72,
-99,74,75,76,77,78,79,-99 };
th(1,1,3); // 16
th(2,0,:); // { 17, 18, 19, 20 };
th(:,1,3); // 8, 16, 24, -99, 40, 48, 56, 64, 72, -99
th(::5 ,:,0:3:2); // 1, 3, 5, 7, 41, 43, 45, 47 ;
If any of the hyperslab arguments collapse to a single value ( a cross-section has been specified), then that dimension is removed from the returned variable. If all the values collapse then a scalar variable is returned
So for example: the following is valid:
th_nw=th(0,:,:) +th(9,:,:);
th_nw has dimensions $lon,$lat
nb the time dim has become degenerate
The following is not valid:
th_nw=th(0,:,0:1) +th(9,:,0:1);
As the $lon now only has two elements. The above can be calculated by using a LHS cast with $lon_nw as replacement dim for $lon.
defdim("lon_nw",2);
th_nw[$lat,$lon_nw]=th(0,:,0:1) +th(9,:,0:1);
Hyperslabs on the Left Hand Side of an assign
When hyperslabing on the LHS ,the expression on the RHS must evaluate to a scalar or a variable/attribute with the same number of elements as the LHS hyperslab
Sets all elements of the last record to zero.
th(9,:,:)=0.0;
Set first element of each lon element to 1.0.
th(:,:,0)=1.0;
Can hyperslab on both sides of an assign.
Sets the last record to the same as the first record
th(9,:,:)=th(0,:,:);
th0 represents pressure at height=0
th1 represents pressure at height=1
Then its possible to hyperslab in the records
P[$time,$height,$lat,$lon]=0.0;
P(:,0,:,:)=th0;
P(:,1,:,:)=th1
Reverse method
If you want to reverse a dimension's elements in an variable use the reverse() method
with at least one dimension argument (this is equivalent to applying a negative stride)
e.g
th_rv=th(1 ,:,:).reverse($lon); // { 12,11,10,9 } ,{16,15,14,13 }
od_rv=od.reverse($time); // {38, 36, 34, 32, 30, 28, 26, 24, 22, 20 }
Permute method
If you want to swap about the dimensions of a variable use the the permute() method. The number and names of dimension arguments must match the dimensions in the variable. If the first dimension in the variable is of record type then this must remain the first dimension. If you want to change the record dimension consider using ncpdq .
Consider the variable:
float three_dmn_var(lat,lev,lon);
three_dmn_var_pm=three_dmn_var.permute($lon,$lat,$lev);
three_dmn_var_pm=
0,4,8,
12,16,20,
1,5,9,
13,17,21,
2,6,10,
14,18,22,
3,7,11,
15,19,23;
Attributes are referred to by var_nm@att_nm
All the following are valid statements
global@text="Test Attributes"; /* Assign a global variable attribute */
a1[$time]=time*20;
a1@long_name="Kelvin";
a1@min=a1.min();
a1@max=a1.max();
a1@min++;
--a1@max; q
a1(0)=a1@min;
a1($time.size-1)=a1@max;
A value list can be used on the RHS of an assign...
a1@trip1={ 1,2,3 } ;
a1@triplet={ a1@min, (a1@min+a1@max)/2, a1@max };
The netcdf specification allows all attribute types to have a size greater than one. The maximum is defined by NC_MAX_ATTRS -The following is an ncdump of the meta-data for variable a1
double a1(time) ;
a1:long_name = "Kelvin" ;
a1:max = 199. ;
a1:min = 21. ;
a1:trip1 = 1, 2, 3 ;
a1:triplet = 21., 110., 199. ;
The size() method can be used with attributes -for example to save an attribute text string in a variable..
defdim("sng_len", a1@long_name.size());
sng_arr[$sng_len]=a1@long_name; // sng_arr now contains "Kelvin"
Attributes defined in a script are stored in memory and are written to Output after script completion. To stop the attribute being written use the ram_delete() method or use a bogus variable name
Attribute Propagation & Inheritance
prs_mdp[time,lat,lon,lev]=P0*hyam+hybm*PS; //prs_mdp get attributes from PO
th_min=1.0 + 2*three_dmn_var_dbl.min($time); //th_min get attributes from three_dmn_var_dbl
If the attribute name contains non-regular characters use ID quoting. See see ID Quoting
'b..m1@c--lost'=23;
The table below lists the postfix character(s) to add to a number literal for type cohesion. To use the new netcdf4 types nco must be compiled/linked to the netcdf4 library and the Output file must be hdf5.
n1[$time]=1UL; // n1 will now by typeNC_UINTn2[$lon]=4b; // n2 will be of typeNC_BYTEn3[$lat]=5ull; // n3 will be of typeNC_UINT64n3@a1=6.0d; // attribute will be typeNC_DOUBLEn3@a2=-666L; // attribute will be typeNC_INT
A floating point number without a postfix will default to NC_DOUBLE. An integer without a postfix will default to type NC_INT. Thre is no postfix for characters. Use a quoted string.
n4[$rlev]=.1 // n4 will be of typeNC_DOUBLEn5[$lon_grd]=2. // n5 will be of typeNC_DOUBLEn6[$gds_crd]=2e3; // n6 will be of typeNC_DOUBLEn6@a1=41; // attribute will be typeNC_INTn6@a2=-21; // attribute will be typeNC_INTn6@units="kelvin" // attribute will be typeNC_CHAR
NC_BYTE a signed 1 byte integer
NC_CHAR a ISO/ASCII character
NC_SHORT a signed 2 byte integer
NC_INT a signed 4 byte integer
NC_FLOAT a single precision floating point number
NC_DOUBLE a double precision floating point number
NC_UBYTE a unsigned 1 byte int
NC_USHORT a unsigned 2-byte int
NC_UINT a unsigned 4-byte int
NC_INT64 a signed 8-byte int
NC_UINT64 a unsigned 8-byte int
The synatax of the if statement is similar to its C counterpart. The Conditional Operator (ternary operator) has also been implemented.
if(exp1)
stmt1;
else if(exp2)
stmt2
else
stmt3
Can use code blocks as well
if(exp1){
stmt1;
stmt1a;
stmt1b;
} else if(exp2)
stmt2
else {
stmt3;
stmt3a;
stmt3b;
}
For a variable or attribute expression to be logically true all its non-missing value elements must be logically true. (i.e non-zero). The expression can be of any type. Unlike C there is no short-circuiting of an expression with the OR (||) AND (&&) operators. The whole expression is evaluated regardless if one of the AND/OR operands are true/false.
A simple example
if(time>0)
print("All values of time are greater than zero\n");
else if( time<0)
print("All values of time are less than zero\n");
else {
time_max=time.max();
time_min=time.min();
print("min value of time=");print(time_min,"%f");
print("max value of time=");print(time_max,"%f");
}
A real example from ddra.nco
if(fl_typ==fl_typ_gcm){
var_nbr_apx=32;
lmn_nbr=1.0*var_nbr_apx*varsz_gcm_4D; /* [nbr] Variable size */
if(nco_op_typ==nco_op_typ_avg){
lmn_nbr_avg=1.0*var_nbr_apx*varsz_gcm_4D; /* [nbr] Averaging block size */
lmn_nbr_wgt=dmnsz_gcm_lat; /* [nbr] Weight size */
} // !nco_op_typ_avg
}else if(fl_typ==fl_typ_stl){
var_nbr_apx=8;
lmn_nbr=1.0*var_nbr_apx*varsz_stl_2D; /* [nbr] Variable size */
if(nco_op_typ==nco_op_typ_avg){
lmn_nbr_avg=1.0*var_nbr_apx*varsz_stl_2D; /* [nbr] Averaging block size */
lmn_nbr_wgt=dmnsz_stl_lat; /* [nbr] Weight size */
} // !nco_op_typ_avg
} // !fl_typ
Conditional Operator
// nb you need netCDF4 to run this example
th_nw=(three_dmn_var_sht >= 0 ? three_dmn_var_sht.uint(): three_dmn_var_sht.int() );
print( variable_name/attribute name/string, format string);
The print function takes a variable name or attribute name or a quoted string and prints the contents in a in a similar fashion to ncks -H .
There is also an optional C style format string argument.
Currently the print function can't print RAM variables or expressions e.g 'print(var_msk*3+4)' is invalid. So if you want to print an expression first assign a variable with the expression; then print the variable.
examples
print(lon);
lon[0]=0
lon[1]=90
lon[2]=180
lon[3]=270
print(lon_2D_rrg,"%3.2f,");
0.00,0.00,180.00,0.00,180.00,0.00,180.00,0.00,
print(mss_val_fst@_FillValue);
mss_val_fst@_FillValue, size = 1 NC_FLOAT, value = -999
print("This function \t is monotonic\n");
This function is monotonic
Missing values operate slightly differently in ncap2 Consider the expression where op is any of the following operators (excluding '=')
Arithmetic operators ( * / % + - ^ )
Binary Operators ( >, >= <, <= ==, !=,==,||,&&, >>,<< )
Assign Operators ( +=,-=,/=, *= )
var1 'op' var2
if var1 has a missing value then this is the value used in the operation else the missing value for var2 is used. if during the element by element operation an element from either operand is equal to the missing value then the missing value is carried through. In this way missing values 'percolate' through an expression.
Missing values associated with Output variables are stored in memory and are written to disk after the script finishes. During script execution its possible (and legal) for the missing value of a variable to take on several different values.
Consider the variable:
int rec_var_int_mss_val_int(time); =-999,2,3,4,5,6,7,8,-999,-999;
rec_var_int_mss_val_int:_FillValue = -999;
n2=rec_var_int_mss_val_int + rec_var_int_mss_val_int.reverse($time);
n2=-999,-999,11,11,11,11,11,11,999,-999;
The following methods are used to edit the missing value associated with a variable. They only work on variables in Output.
set_miss(expr)change_miss(expr)get_miss()delete_miss()
th=three_dmn_var_dbl;
th.change_miss(-1e10d);
/* set values less than 0 or greater than 50 to missing value */
where( th <0.0 || th > 50.0)
th=th.get_miss();
Another example
new[$time,$lat,$lon]=1.0;
new.set_miss(-997.0);
/* extract only elements divisible by 3 */
where ( three_dmn_var_dbl%3 == 0 )
new=three_dmn_var_dbl;
elsewhere
new=new.get_miss();
The convention within this document is that methods can be used as
functions.
However, functions are not and cannot be used as methods.
Methods can be daisy changed together and their synatax is cleaner than
functions.
Method names are reserved words and CANNOT be used as variable names.
The command ncap2 -f shows the complete list of methods available
on your build.
n2=sin(theta) or n2=theta.sin()
n2=sin(theta)^2 +cos(theta)^2 or n2=theta.sin().pow(2) + theta.cos()^2
The below statement converts three_dmn_var_sht to type double, finds the average, then converts this average back to type short.
three_avg=three_dmn_var_sht.double().avg().short();
Aggregate Methods
avg()sqravg()avgsqr()max()min()rms()rmssdn()ttl() or total() // Average a variable over time
four_time_avg=four_dmn_rec_var($time);
Packing Methods
pack() & pack_short()NC_SHORT
pack_byte()NC_BYTE
pack_short()NC_SHORT
pack_int()NC_INT
unpack()Basic Methods
These methods work with variables and attributes. They have no arguments
size()ndims()type()Utility Methods
set_miss(expr)change_miss(expr)get_miss()delete_miss()ram_write()ram_delete()PDQ Methods
reverse(dim args)permute(dim args) lat_2D_rrg_new=lat_2D_rrg.permute($lon,$lat).reverse($lon);
lat_2D_rrg_new=0,90,-30,30,-30,30,-90,0
Type Conversion Methods
byte()NC_BYTE a signed 1 byte integer
char()NC_CHAR a ISO/ASCII character
short()NC_SHORT a signed 2 byte integer
int()NC_INT a signed 4 byte integer
float()NC_FLOAT a single precision floating point number
double()NC_DOUBLE a double precision floating point number
ubyte()NC_UBYTE a unsigned 1 byte int
ushort()NC_USHORT a unsigned 2-byte int
uint()NC_UINT a unsigned 4-byte int
int64()NC_INT64 a signed 8-byte int
uint64()NC_UINT64 a unsigned 8-byte int *
Intrinsic Mathematical Methods
The list of mathematical methods is system dependant.
For the full list see Intrinsic mathematical methods
All the mathematical methods take a single operand ,with the exception of atan2 and pow which take two.
If the operand type is less than float then the result will be of type float. If the operand is type double then the result will be type double. Like the other methods, you are free to use the mathematical methods as functions.
n1=pow(2,3.0f) // n1 type float
n2=atan2(2,3.0) // n2 type double
n3=1/(three_dmn_var_dbl.cos().pow(2))-tan(three_dmn_var_dbl)^2; // n3 type double
RAM variables are used in place of regular variables to speed things up. For example in a loop or where a variable is very frequently referenced. To declare and define a RAM variable simply prefix the variable name with * when the variable is declared/initialized.
To delete a RAM variable (recover some memory) use the ram_delete() method. To convert a RAM variable to a regular disk variable in output use ram_write() method.
The following is valid:
*temp[$time,$lat,lon]=10.0; // Cast
*temp_avg=temp.avg($time); // Regular assign
....
temp.ram_delete(); // Delete RAM variable
temp_avg.ram_write(); // Write Variable to output
Other Assigns
// Create a RAM variable from the variable "one" in Input and increment its elements
*one++;
// Create a RAM variable from the variable three in Input and multiply its contents by 10
// Create a RAM variable from the variable four in Input and then add the variable "three" to
// its contents.
*four+=*three*=10; // three=30, four=34
A where() combines the definition and application of a mask all in one go and can lead to succinct code.
The full syntax of a where() statement is as follows:
// Single assign (nb the else block is optional)
where (mask)
var1=expr1;
elsewhere
var1=expr2;
// Multiple assigns
where( mask) {
var1=expr1;
var2=expr2;
...
} elsewhere {
var1=expr3
var2=expr4
var3=expr5;
...
}
example:
Consider the variables:
float lon_2D_rct(lat,lon);
float var_msk(lat,lon);
Suppose we want to multiply by two the elements for which var_msk is equal to 1;
where(var_msk==1)
lon_2D_rct=2*lon_2D_rct;
Another example
Suppose we have the variable
int RDM(time);
And we want to set the values less than 8 or greater than 80 to 0.
where(RDM <8 || RDM >80)
RDM=0;
A more complex example.
Consider the situation where we have irregularly gridded data, described using rank 2 variables:
double lat(south_north,east_west)
To find the average temperature in a region [lat_min,lat_max] and [lon_min,lon_max]:
double lon(south_north,east_west)
double temperature(south_north,east west)
temperature_msk[$south_north,$east_west]=0.0;
where(lat >= lat_min && lat <= lat_max) && (lon >= lon_min && > lon <= lon_max)
temperature_msk=temperature;
elsewhere
temperature_msk=temperature@_FillValue;
temp_avg=temperature_msk.avg();
temp_max=temperature.max();
In ncap there are for() loops and while() loops. They are currently completely unoptimized So use them with RAM variables unless you want thrash your disk to death. To break out of a loop use "break" command. To iterate to the next cycle use the "continue" command.
// Follwing sets elements in variable double temp(time,lat)
// If element < 0 set to 0, if element >100 set to 100
*sz_idx=$time.size;
*sz_jdx=$lat.size;
for(*idx=0 ; idx<sz_idx ; idx++)
for(*jdx=0 ; jdx<sz_jdx; jdx++)
if( temp(idx,jdx) >100 ) temp(idx,jdx)=100.0;
else if( temp(idx,jdx) <0 ) temp(idx,jdx)=0.0;
// See if values of of a co-ordinate variable double lat(lat) are monotonic
*sz=$lat.size;
for(*idx=1 ; idx<sz;idx++)
if( lat(idx)-lat(idx-1) < 0.0)
break;
if(idx==sz)
print("lat co-ordinate is monotonic\n");
else
print("lat co-ordinate is NOT monotonic\n");
// Sum odd elements
*idx=0;
*sz=$lat_nw.size;
*sum=0.0;
while(idx<sz){
if( lat(idx) % 2) sum+=lat(idx);
idx++;
}
ram_write(sum);
print("Total of odd elements ");print(sum);print("\n");
The synatax of an include file is:
#include "script"
The script filename is searched relative to the run directory. Its possible to nest include files to an arbitrary depth. A handy use of inlcude files is to store often used constants. Use RAM variables of you don't want these constants written to Output.
*pi=3.1415926535;
*h=6.62607095e-34;
e=2.71828;
In ncap there are two ways to sort. The first is a regular sort. This sorts ALL the elements of a variable or attribute without regard to any dimensions. The second method applies a sort map to a variable. To apply this sort map the size of the variable must be exactly divisible by the size of the sort map. The method sort(var_in,&var_map) is overloaded. The second optional argument is a call_by_ref variable which will hold the sort map.
a1[$time]={10,2,3,4,6,5,7,3,4,1};
a1_sort=sort(a1);
print(a1_sort);
// 1, 2, 3, 3, 4, 4, 5, 6, 7, 10 ;
a2[$lon]={2,1,4,3};
a2_sort=sort(a2,&a2_map);
print(a2);
// 1, 2, 3, 4
print(a2_map);
// 1, 0, 3, 2 ;
If the map variable doesn't exist prior to the sort call, then it will be created with the same shape as the input variable and be of type NC_INT. If the map variable already exists, then the only restriction is that it be of at least the same size as the input variable. To apply a map use dsort(var_in,var_map).
defdim("nlat",5);
a3[$lon]={2,5,3,7};
a4[$nlat,$lon]={
1, 2, 3, 4,
5, 6, 7, 8,
9,10,11,12,
13,14,15,16,
17,18,19,20};
a3_sort=sort(a3,&a3_map);
print(a3_map);
// 0, 2, 1, 3 ;
a5_sort=dsort(a5,a3_map);
print(a5_sort);
// 1, 3, 2, 4,
// 5, 7, 6, 8,
// 9,11,10,12,
// 13,15,14,16,
// 17,19,18,20 ;
a3_map2[$nlat]={4,3,0,2,1 };
a5_sort2=dsort(a5,a3_map2);
print(a5_sort2);
// 3, 5, 4, 2, 1
// 8, 10, 9,7, 6,
// 13,15,14,12,11,
// 18,20,19,17,16
As in the above example you a free to create your own mask. If you wish to sort in decending order then use the reverse() method after the sort.
NCO is capable of analyzing datasets for many different underlying coordinate grid types. netCDF was developed for and initially used with grids comprised of orthogonal dimensions forming a rectangular coordinate system. We call such grids standard grids. It is increasingly common for datasets to use metadata to describe much more complex grids. Let us first define three important coordinate grid properties: rectangularity, regularity, and fxm.
Grids are regular if the spacing between adjacent is constant. For example, a 4-by-5 degree latitude-longitude grid is regular because the spacings between adjacent latitudes (4 degrees) are constant as are the (5 degrees) spacings between adjacent longitudes. Spacing in irregular grids depends on the location along the coordinate. Grids such as Gaussian grids have uneven spacing in latitude (points cluster near the equator) and so are irregular.
Grids are rectangular if the number of elements in any dimension is not a function of any other dimension. For example, a T42 Gaussian latitude-longitude grid is rectangular because there are the same number of longitudes (128) for each of the (64) latitudes. Grids are non-rectangular if the elements in any dimension depend on another dimension. Non-rectangular grids present many special challenges to analysis software like NCO.
Wrapped coordinates (see Wrapped Coordinates), such as longitude, are independent of these grid properties (regularity, rectangularity).
The preferred NCO technique to analyze data on non-standard coordinate grids is to create a region mask with ncap2, and then to use the mask within ncap2 for variable-specific processing, and/or with other operators (e.g., ncwa, ncdiff) for entire file processing.
Before describing the construction of masks, let us review how irregularly gridded geoscience data are described. Say that latitude and longitude are stored as R-dimensional arrays and the product of the dimension sizes is the total number of elements N in the other variables. Geoscience applications tend to use R=1, R=2, and R=3.
If the grid is has no simple representation (e.g., discontinuous) then it makes sense to store all coordinates as 1D arrays with the same size as the number of grid points. These gridpoints can be completely independent of all the other (own weight, area, etc.).
R=1: lat(number_of_gridpoints) and lon(number_of_gridpoints)
If the horizontal grid is time-invariant then R=2 is common:
R=2: lat(south_north,east_west) and lon(south_north,east_west)
The WRF (Weather and Research Forecast) model uses R=3
R=3: lat(time,south_north,east_west), lon(time,south_north,east_west)
and so supports grids that change with time.
Grids with R > 1 often use missing values to indicated empty points. For example, so-called "staggered grids" will use fewer east_west points near the poles and more near the equator. netCDF only accepts rectangular arrays so space must be allocated for the maximum number of east_west points at all latitudes. Then the application writes missing values into the unused points near the poles.
Let's demonstrate the recommended ncap2 analysis technique by constructing a region mask for an R=2 grid. We wish to find, say, the mean temperature within [lat_min,lat_max] and [lon_min,lon_max]:
ncap2 -s 'mask= (lat >= lat_min && lat <= lat_max) && \
(lon >= lon_min && lon <= lon_max);' in.nc out.nc
Once you have a mask, you can use it on specific variables:
ncap2 -s 'temperature_avg=(temperature*mask).avg()' in.nc out.nc
and you can apply it to entire files:
ncwa -a lat,lon -m mask -w area in.nc out.nc
You can put this altogether on the command line or in a script, e.g., cleaner.
cat > ncap2.in << EOF
mask = (lat >= lat_min && lat <= lat_max) && (lon >= lon_min && > lon <= lon_max);
if(mask.total() > 0){ // Check that mask contains some valid values
temperature_avg=(temperature*mask).avg(); // Average temperature
temperature_max=(temperature*mask).max(); // Maximum temperature
}
EOF
ncap2 -S ncap2.in in.nc out.nc
For the WRF file creating the mask looks like
mask = (XLAT >= lat_min && XLAT <= lat_max) && (XLONG >= lon_min && > XLONG <= lon_max);
In practice with WRF it's a bit more complicated because you must use the global metadata to determine the grid staggering and offsets to translate XLAT and XLONG into real latitudes and longitudes and missing points. The WRF grid documentation should describe this.
A few notes: Irregular regions are the union of arrays lat/lon_min/max's. The mask procedure is identical for all R.
As of version 4.0.0 NCO has internal routines to perform bilinear interpolation on gridded data sets.
In mathematics, bilinear interpolation is an extension of linear interpolation for interpolating functions of two variables on a regular grid. The idea is to perform linear interpolation first in one direction, and then again in the other direction.
Suppose we have an irregular grid of data temperature[lat,lon], with co-ordinate vars lat[lat], lon[lon]. And we wish to find the temperature at an arbitary point [X,Y] within the grid. If we can locate lat_min,lat_max and lon_min,lon_max such that lat_min <= X <= lat_max and lon_min <=Y <=lon_max then we can interpolate in two dimensions to find the temperature at [X,Y].
the general form of the ncap interpolation function is as follows:-
var_out=bilinear_interp(grid_in, grid_out, grid_out_x, grid_out_y, grid_in_x, grid_in_y)
grid_ingrid_in_x.size()*grid_in_y*size()
grid_outgrid_out_x.size()*grid_out_y*size()grid_out_xgrid_out_ygrid_in_xgrid_in_yPrior to calculations all arguments are converted to type NC_DOUBLE. After calculations var_out is converted to the input type of grid_in.
Suppose the first part of an ncap2 script is:
/****************************************/
defdim("X",4);
defdim("Y",5);
//Temperature
T_in[$X,$Y]=
{100, 200, 300, 400, 500,
101, 202, 303, 404, 505,
102, 204, 306, 408, 510,
103, 206, 309, 412, 515.0 };
//Co-ordinate Vars
x_in[$X]={ 0.0,1.0,2.0,3.01 };
y_in[$Y]={ 1.0,2.0,3,4,5 };
/***************************************/
Now we interpolate with the following variables:
/***************************************/
defdim("Xn",3);
defdim("Yn",4);
T_out[$Xn,$Yn]=0.0;
x_out[$Xn]={0.0,0.02,3.01 };
y_out[$Yn]={1.1,2.0,3,4 };
var_out=bilinear_interp(T_in,T_out,x_out,y_out,x_in,y_in);
print(var_out);
// 110, 200, 300, 400,
// 110.022, 200.04, 300.06, 400.08,
// 113.3, 206, 309, 412 ;
/***************************************/
Its possible to use the call to interpolate a single point:
/***************************************/
var_out=bilinear_interp(T_in,0.0,3.0,4.99,x_in,y_in);
print(var_out);
// 513.920594059406
/***************************************/
As of version 3.9.6 (released January, 2009), NCO can link to the GNU Scientific Library (GSL). ncap can access most GSL special functions including Airy, Bessel, error, gamma, beta, hypergeometric, and Legendre functions and elliptical integrals. GSL must be version 1.4 or later. To list the GSL functions available with your NCO build, use ncap2 -f | grep ^gsl.
The function names used by ncap2 mirror their GSL names. The NCO wrappers for GSL functions automatically call the error-handling version of the GSL function when available 32. This allows NCO to return a missing value when the GSL library encounters a domain error or a floating point exception. The slow-down due to calling the error-handling version of the GSL numerical functions was found to be negligible (please let us know if you find otherwise).
Consider the gamma function.
The GSL function prototype is
int gsl_sf_gamma_e(const double x, gsl_sf_result * result)
The ncap script would be:
lon_in[lon]={-1,0.1,0,2,0.3};
lon_out=gsl_sf_gamma(lon_in);
lon_out= _, 9.5135, 4.5908, 2.9915
The first value is set to _FillValue since the gamma
function is undefined for negative integers.
If the input variable has a missing value then this value is used.
Otherwise, the default double fill value is used
(defined in the netCDF header netcdf.h as
NC_FILL_DOUBLE = 9.969e+36).
Consider a call to a Bessel function with GSL
prototype
int gsl_sf_bessel_Jn_e(int n, double x, gsl_sf_result * result)
An ncap script would be
lon_out=gsl_sf_bessel_Jn(2,lon_in);
lon_out=0.11490, 0.0012, 0.00498, 0.011165
This computes the Bessel function of order n=2 for every value in
lon_in.
The Bessel order argument, an integer, can also be a non-scalar
variable, i.e., an array.
n_in[lon]={0,1,2,3};
lon_out=gsl_sf_bessel_Jn(n_in,0.5);
lon_out= 0.93846, 0.24226, 0.03060, 0.00256
Arguments to GSL wrapper functions in ncap
must conform to one another, i.e., they must share the same sub-set of
dimensions.
For example: three_out=gsl_sf_bessel_Jn(n_in,three_dmn_var_dbl)
is valid because the variable three_dmn_var_dbl has a lon
dimension, so n_in in can be broadcast to conform to
three_dmn_var_dbl.
However time_out=gsl_sf_bessel_Jn(n_in,time) is invalid.
Consider the elliptical integral with prototype
int gsl_sf_ellint_RD_e(double x, double y, double z, gsl_mode_t mode, gsl_sf_result * result)
three_out=gsl_sf_ellint_RD(0.5,time,three_dmn_var_dbl);
The three arguments are all conformable so the above ncap call is valid. The mode argument in the function prototype controls the convergence of the algorithm. It also appears in the Airy Function prototypes. It can be set by defining the environment variable GSL_PREC_MODE. If unset it defaults to the value GSL_PREC_DOUBLE. See the GSL manual for more details.
export GSL_PREC_MODE=0 // GSL_PREC_DOUBLE
export GSL_PREC_MODE=1 // GSL_PREC_SINGLE
export GSL_PREC_MODE=2 // GSL_PREC_APPROX
The ncap wrappers to the array functions are slightly different. Lets consider the following gsl prototype
int gsl_sf_bessel_Jn_array(int nmin, int nmax, double x, double *result_array)
b1=lon.double();
x=0.5;
status=gsl_sf_bessel_Jn_array(1,4,x,&b1);
print(status);
b1=0.24226, 0.0306, 0.00256, 0.00016 ;
This calculates the bessel function of x=0.5 for n=1 to 4. The first three arguments are scalar values. if a non-scalar variable is supplied as an argument then only the first value is used.The final argument is the variable where the results go ( note the '&' this indicates a call by reference). This final argument must be of type double and must be of least size (nmax-nmin+1). If either of these conditions are not met then then the function will blow out with an error message. The function/wrapper returns a status flag. Zero indicates success.
Lets look at another array function
int gsl_sf_legendre_Pl_array( int lmax, double x, double *result_array);
a1=time.double();
x=0.3;
status=gsl_sf_legendre_Pl_array(a1.size()-1, x,&a1);
print(status);
This call calculates P_l(0.3) for l=0..9. Note |x|<=1, otherwise there will be a domain error. See the GSL documentation for more details.
Below is table detailing what GSL functions have been implemented. This table is correct for GSL version 1.10. To see what functions are available on your build run the command ncap2 -f |grep ^gsl . To see this table along with the GSL C function prototypes look at the spreadsheet doc/nco_gsl.ods.
| GSL NAME | I | NCAP FUNCTION CALL
|
| gsl_sf_airy_Ai_e | Y | gsl_sf_airy_Ai(dbl_expr)
|
| gsl_sf_airy_Bi_e | Y | gsl_sf_airy_Bi(dbl_expr)
|
| gsl_sf_airy_Ai_scaled_e | Y | gsl_sf_airy_Ai_scaled(dbl_expr)
|
| gsl_sf_airy_Bi_scaled_e | Y | gsl_sf_airy_Bi_scaled(dbl_expr)
|
| gsl_sf_airy_Ai_deriv_e | Y | gsl_sf_airy_Ai_deriv(dbl_expr)
|
| gsl_sf_airy_Bi_deriv_e | Y | gsl_sf_airy_Bi_deriv(dbl_expr)
|
| gsl_sf_airy_Ai_deriv_scaled_e | Y | gsl_sf_airy_Ai_deriv_scaled(dbl_expr)
|
| gsl_sf_airy_Bi_deriv_scaled_e | Y | gsl_sf_airy_Bi_deriv_scaled(dbl_expr)
|
| gsl_sf_airy_zero_Ai_e | Y | gsl_sf_airy_zero_Ai(uint_expr)
|
| gsl_sf_airy_zero_Bi_e | Y | gsl_sf_airy_zero_Bi(uint_expr)
|
| gsl_sf_airy_zero_Ai_deriv_e | Y | gsl_sf_airy_zero_Ai_deriv(uint_expr)
|
| gsl_sf_airy_zero_Bi_deriv_e | Y | gsl_sf_airy_zero_Bi_deriv(uint_expr)
|
| gsl_sf_bessel_J0_e | Y | gsl_sf_bessel_J0(dbl_expr)
|
| gsl_sf_bessel_J1_e | Y | gsl_sf_bessel_J1(dbl_expr)
|
| gsl_sf_bessel_Jn_e | Y | gsl_sf_bessel_Jn(int_expr,dbl_expr)
|
| gsl_sf_bessel_Jn_array | Y | status=gsl_sf_bessel_Jn_array(int,int,double,&var_out)
|
| gsl_sf_bessel_Y0_e | Y | gsl_sf_bessel_Y0(dbl_expr)
|
| gsl_sf_bessel_Y1_e | Y | gsl_sf_bessel_Y1(dbl_expr)
|
| gsl_sf_bessel_Yn_e | Y | gsl_sf_bessel_Yn(int_expr,dbl_expr)
|
| gsl_sf_bessel_Yn_array | Y | gsl_sf_bessel_Yn_array
|
| gsl_sf_bessel_I0_e | Y | gsl_sf_bessel_I0(dbl_expr)
|
| gsl_sf_bessel_I1_e | Y | gsl_sf_bessel_I1(dbl_expr)
|
| gsl_sf_bessel_In_e | Y | gsl_sf_bessel_In(int_expr,dbl_expr)
|
| gsl_sf_bessel_In_array | Y | status=gsl_sf_bessel_In_array(int,int,double,&var_out)
|
| gsl_sf_bessel_I0_scaled_e | Y | gsl_sf_bessel_I0_scaled(dbl_expr)
|
| gsl_sf_bessel_I1_scaled_e | Y | gsl_sf_bessel_I1_scaled(dbl_expr)
|
| gsl_sf_bessel_In_scaled_e | Y | gsl_sf_bessel_In_scaled(int_expr,dbl_expr)
|
| gsl_sf_bessel_In_scaled_array | Y | staus=gsl_sf_bessel_In_scaled_array(int,int,double,&var_out)
|
| gsl_sf_bessel_K0_e | Y | gsl_sf_bessel_K0(dbl_expr)
|
| gsl_sf_bessel_K1_e | Y | gsl_sf_bessel_K1(dbl_expr)
|
| gsl_sf_bessel_Kn_e | Y | gsl_sf_bessel_Kn(int_expr,dbl_expr)
|
| gsl_sf_bessel_Kn_array | Y | status=gsl_sf_bessel_Kn_array(int,int,double,&var_out)
|
| gsl_sf_bessel_K0_scaled_e | Y | gsl_sf_bessel_K0_scaled(dbl_expr)
|
| gsl_sf_bessel_K1_scaled_e | Y | gsl_sf_bessel_K1_scaled(dbl_expr)
|
| gsl_sf_bessel_Kn_scaled_e | Y | gsl_sf_bessel_Kn_scaled(int_expr,dbl_expr)
|
| gsl_sf_bessel_Kn_scaled_array | Y | status=gsl_sf_bessel_Kn_scaled_array(int,int,double,&var_out)
|
| gsl_sf_bessel_j0_e | Y | gsl_sf_bessel_J0(dbl_expr)
|
| gsl_sf_bessel_j1_e | Y | gsl_sf_bessel_J1(dbl_expr)
|
| gsl_sf_bessel_j2_e | Y | gsl_sf_bessel_j2(dbl_expr)
|
| gsl_sf_bessel_jl_e | Y | gsl_sf_bessel_jl(int_expr,dbl_expr)
|
| gsl_sf_bessel_jl_array | Y | status=gsl_sf_bessel_jl_array(int,double,&var_out)
|
| gsl_sf_bessel_jl_steed_array | Y | gsl_sf_bessel_jl_steed_array
|
| gsl_sf_bessel_y0_e | Y | gsl_sf_bessel_Y0(dbl_expr)
|
| gsl_sf_bessel_y1_e | Y | gsl_sf_bessel_Y1(dbl_expr)
|
| gsl_sf_bessel_y2_e | Y | gsl_sf_bessel_y2(dbl_expr)
|
| gsl_sf_bessel_yl_e | Y | gsl_sf_bessel_yl(int_expr,dbl_expr)
|
| gsl_sf_bessel_yl_array | Y | status=gsl_sf_bessel_yl_array(int,double,&var_out)
|
| gsl_sf_bessel_i0_scaled_e | Y | gsl_sf_bessel_I0_scaled(dbl_expr)
|
| gsl_sf_bessel_i1_scaled_e | Y | gsl_sf_bessel_I1_scaled(dbl_expr)
|
| gsl_sf_bessel_i2_scaled_e | Y | gsl_sf_bessel_i2_scaled(dbl_expr)
|
| gsl_sf_bessel_il_scaled_e | Y | gsl_sf_bessel_il_scaled(int_expr,dbl_expr)
|
| gsl_sf_bessel_il_scaled_array | Y | status=gsl_sf_bessel_il_scaled_array(int,double,&var_out)
|
| gsl_sf_bessel_k0_scaled_e | Y | gsl_sf_bessel_K0_scaled(dbl_expr)
|
| gsl_sf_bessel_k1_scaled_e | Y | gsl_sf_bessel_K1_scaled(dbl_expr)
|
| gsl_sf_bessel_k2_scaled_e | Y | gsl_sf_bessel_k2_scaled(dbl_expr)
|
| gsl_sf_bessel_kl_scaled_e | Y | gsl_sf_bessel_kl_scaled(int_expr,dbl_expr)
|
| gsl_sf_bessel_kl_scaled_array | Y | status=gsl_sf_bessel_kl_scaled_array(int,double,&var_out)
|
| gsl_sf_bessel_Jnu_e | Y | gsl_sf_bessel_Jnu(dbl_expr,dbl_expr)
|
| gsl_sf_bessel_Ynu_e | Y | gsl_sf_bessel_Ynu(dbl_expr,dbl_expr)
|
| gsl_sf_bessel_sequence_Jnu_e | N | gsl_sf_bessel_sequence_Jnu
|
| gsl_sf_bessel_Inu_scaled_e | Y | gsl_sf_bessel_Inu_scaled(dbl_expr,dbl_expr)
|
| gsl_sf_bessel_Inu_e | Y | gsl_sf_bessel_Inu(dbl_expr,dbl_expr)
|
| gsl_sf_bessel_Knu_scaled_e | Y | gsl_sf_bessel_Knu_scaled(dbl_expr,dbl_expr)
|
| gsl_sf_bessel_Knu_e | Y | gsl_sf_bessel_Knu(dbl_expr,dbl_expr)
|
| gsl_sf_bessel_lnKnu_e | Y | gsl_sf_bessel_lnKnu(dbl_expr,dbl_expr)
|
| gsl_sf_bessel_zero_J0_e | Y | gsl_sf_bessel_zero_J0(uint_expr)
|
| gsl_sf_bessel_zero_J1_e | Y | gsl_sf_bessel_zero_J1(uint_expr)
|
| gsl_sf_bessel_zero_Jnu_e | N | gsl_sf_bessel_zero_Jnu
|
| gsl_sf_clausen_e | Y | gsl_sf_clausen(dbl_expr)
|
| gsl_sf_hydrogenicR_1_e | N | gsl_sf_hydrogenicR_1
|
| gsl_sf_hydrogenicR_e | N | gsl_sf_hydrogenicR
|
| gsl_sf_coulomb_wave_FG_e | N | gsl_sf_coulomb_wave_FG
|
| gsl_sf_coulomb_wave_F_array | N | gsl_sf_coulomb_wave_F_array
|
| gsl_sf_coulomb_wave_FG_array | N | gsl_sf_coulomb_wave_FG_array
|
| gsl_sf_coulomb_wave_FGp_array | N | gsl_sf_coulomb_wave_FGp_array
|
| gsl_sf_coulomb_wave_sphF_array | N | gsl_sf_coulomb_wave_sphF_array
|
| gsl_sf_coulomb_CL_e | N | gsl_sf_coulomb_CL
|
| gsl_sf_coulomb_CL_array | N | gsl_sf_coulomb_CL_array
|
| gsl_sf_coupling_3j_e | N | gsl_sf_coupling_3j
|
| gsl_sf_coupling_6j_e | N | gsl_sf_coupling_6j
|
| gsl_sf_coupling_RacahW_e | N | gsl_sf_coupling_RacahW
|
| gsl_sf_coupling_9j_e | N | gsl_sf_coupling_9j
|
| gsl_sf_coupling_6j_INCORRECT_e | N | gsl_sf_coupling_6j_INCORRECT
|
| gsl_sf_dawson_e | Y | gsl_sf_dawson(dbl_expr)
|
| gsl_sf_debye_1_e | Y | gsl_sf_debye_1(dbl_expr)
|
| gsl_sf_debye_2_e | Y | gsl_sf_debye_2(dbl_expr)
|
| gsl_sf_debye_3_e | Y | gsl_sf_debye_3(dbl_expr)
|
| gsl_sf_debye_4_e | Y | gsl_sf_debye_4(dbl_expr)
|
| gsl_sf_debye_5_e | Y | gsl_sf_debye_5(dbl_expr)
|
| gsl_sf_debye_6_e | Y | gsl_sf_debye_6(dbl_expr)
|
| gsl_sf_dilog_e | N | gsl_sf_dilog
|
| gsl_sf_complex_dilog_xy_e | N | gsl_sf_complex_dilog_xy_e
|
| gsl_sf_complex_dilog_e | N | gsl_sf_complex_dilog
|
| gsl_sf_complex_spence_xy_e | N | gsl_sf_complex_spence_xy_e
|
| gsl_sf_multiply_e | N | gsl_sf_multiply
|
| gsl_sf_multiply_err_e | N | gsl_sf_multiply_err
|
| gsl_sf_ellint_Kcomp_e | Y | gsl_sf_ellint_Kcomp(dbl_expr)
|
| gsl_sf_ellint_Ecomp_e | Y | gsl_sf_ellint_Ecomp(dbl_expr)
|
| gsl_sf_ellint_Pcomp_e | Y | gsl_sf_ellint_Pcomp(dbl_expr,dbl_expr)
|
| gsl_sf_ellint_Dcomp_e | Y | gsl_sf_ellint_Dcomp(dbl_expr)
|
| gsl_sf_ellint_F_e | Y | gsl_sf_ellint_F(dbl_expr,dbl_expr)
|
| gsl_sf_ellint_E_e | Y | gsl_sf_ellint_E(dbl_expr,dbl_expr)
|
| gsl_sf_ellint_P_e | Y | gsl_sf_ellint_P(dbl_expr,dbl_expr,dbl_expr)
|
| gsl_sf_ellint_D_e | Y | gsl_sf_ellint_D(dbl_expr,dbl_expr,dbl_expr)
|
| gsl_sf_ellint_RC_e | Y | gsl_sf_ellint_RC(dbl_expr,dbl_expr)
|
| gsl_sf_ellint_RD_e | Y | gsl_sf_ellint_RD(dbl_expr,dbl_expr,dbl_expr)
|
| gsl_sf_ellint_RF_e | Y | gsl_sf_ellint_RF(dbl_expr,dbl_expr,dbl_expr)
|
| gsl_sf_ellint_RJ_e | Y | gsl_sf_ellint_RJ(dbl_expr,dbl_expr,dbl_expr,dbl_expr)
|
| gsl_sf_elljac_e | N | gsl_sf_elljac
|
| gsl_sf_erfc_e | Y | gsl_sf_erfc(dbl_expr)
|
| gsl_sf_log_erfc_e | Y | gsl_sf_log_erfc(dbl_expr)
|
| gsl_sf_erf_e | Y | gsl_sf_erf(dbl_expr)
|
| gsl_sf_erf_Z_e | Y | gsl_sf_erf_Z(dbl_expr)
|
| gsl_sf_erf_Q_e | Y | gsl_sf_erf_Q(dbl_expr)
|
| gsl_sf_hazard_e | Y | gsl_sf_hazard(dbl_expr)
|
| gsl_sf_exp_e | Y | gsl_sf_exp(dbl_expr)
|
| gsl_sf_exp_e10_e | N | gsl_sf_exp_e10
|
| gsl_sf_exp_mult_e | Y | gsl_sf_exp_mult(dbl_expr,dbl_expr)
|
| gsl_sf_exp_mult_e10_e | N | gsl_sf_exp_mult_e10
|
| gsl_sf_expm1_e | Y | gsl_sf_expm1(dbl_expr)
|
| gsl_sf_exprel_e | Y | gsl_sf_exprel(dbl_expr)
|
| gsl_sf_exprel_2_e | Y | gsl_sf_exprel_2(dbl_expr)
|
| gsl_sf_exprel_n_e | Y | gsl_sf_exprel_n(int_expr,dbl_expr)
|
| gsl_sf_exp_err_e | Y | gsl_sf_exp_err(dbl_expr,dbl_expr)
|
| gsl_sf_exp_err_e10_e | N | gsl_sf_exp_err_e10
|
| gsl_sf_exp_mult_err_e | N | gsl_sf_exp_mult_err
|
| gsl_sf_exp_mult_err_e10_e | N | gsl_sf_exp_mult_err_e10
|
| gsl_sf_expint_E1_e | Y | gsl_sf_expint_E1(dbl_expr)
|
| gsl_sf_expint_E2_e | Y | gsl_sf_expint_E2(dbl_expr)
|
| gsl_sf_expint_En_e | Y | gsl_sf_expint_En(int_expr,dbl_expr)
|
| gsl_sf_expint_E1_scaled_e | Y | gsl_sf_expint_E1_scaled(dbl_expr)
|
| gsl_sf_expint_E2_scaled_e | Y | gsl_sf_expint_E2_scaled(dbl_expr)
|
| gsl_sf_expint_En_scaled_e | Y | gsl_sf_expint_En_scaled(int_expr,dbl_expr)
|
| gsl_sf_expint_Ei_e | Y | gsl_sf_expint_Ei(dbl_expr)
|
| gsl_sf_expint_Ei_scaled_e | Y | gsl_sf_expint_Ei_scaled(dbl_expr)
|
| gsl_sf_Shi_e | Y | gsl_sf_Shi(dbl_expr)
|
| gsl_sf_Chi_e | Y | gsl_sf_Chi(dbl_expr)
|
| gsl_sf_expint_3_e | Y | gsl_sf_expint_3(dbl_expr)
|
| gsl_sf_Si_e | Y | gsl_sf_Si(dbl_expr)
|
| gsl_sf_Ci_e | Y | gsl_sf_Ci(dbl_expr)
|
| gsl_sf_atanint_e | Y | gsl_sf_atanint(dbl_expr)
|
| gsl_sf_fermi_dirac_m1_e | Y | gsl_sf_fermi_dirac_m1(dbl_expr)
|
| gsl_sf_fermi_dirac_0_e | Y | gsl_sf_fermi_dirac_0(dbl_expr)
|
| gsl_sf_fermi_dirac_1_e | Y | gsl_sf_fermi_dirac_1(dbl_expr)
|
| gsl_sf_fermi_dirac_2_e | Y | gsl_sf_fermi_dirac_2(dbl_expr)
|
| gsl_sf_fermi_dirac_int_e | Y | gsl_sf_fermi_dirac_int(int_expr,dbl_expr)
|
| gsl_sf_fermi_dirac_mhalf_e | Y | gsl_sf_fermi_dirac_mhalf(dbl_expr)
|
| gsl_sf_fermi_dirac_half_e | Y | gsl_sf_fermi_dirac_half(dbl_expr)
|
| gsl_sf_fermi_dirac_3half_e | Y | gsl_sf_fermi_dirac_3half(dbl_expr)
|
| gsl_sf_fermi_dirac_inc_0_e | Y | gsl_sf_fermi_dirac_inc_0(dbl_expr,dbl_expr)
|
| gsl_sf_lngamma_e | Y | gsl_sf_lngamma(dbl_expr)
|
| gsl_sf_lngamma_sgn_e | N | gsl_sf_lngamma_sgn
|
| gsl_sf_gamma_e | Y | gsl_sf_gamma(dbl_expr)
|
| gsl_sf_gammastar_e | Y | gsl_sf_gammastar(dbl_expr)
|
| gsl_sf_gammainv_e | Y | gsl_sf_gammainv(dbl_expr)
|
| gsl_sf_lngamma_complex_e | N | gsl_sf_lngamma_complex
|
| gsl_sf_taylorcoeff_e | Y | gsl_sf_taylorcoeff(int_expr,dbl_expr)
|
| gsl_sf_fact_e | Y | gsl_sf_fact(uint_expr)
|
| gsl_sf_doublefact_e | Y | gsl_sf_doublefact(uint_expr)
|
| gsl_sf_lnfact_e | Y | gsl_sf_lnfact(uint_expr)
|
| gsl_sf_lndoublefact_e | Y | gsl_sf_lndoublefact(uint_expr)
|
| gsl_sf_lnchoose_e | N | gsl_sf_lnchoose
|
| gsl_sf_choose_e | N | gsl_sf_choose
|
| gsl_sf_lnpoch_e | Y | gsl_sf_lnpoch(dbl_expr,dbl_expr)
|
| gsl_sf_lnpoch_sgn_e | N | gsl_sf_lnpoch_sgn
|
| gsl_sf_poch_e | Y | gsl_sf_poch(dbl_expr,dbl_expr)
|
| gsl_sf_pochrel_e | Y | gsl_sf_pochrel(dbl_expr,dbl_expr)
|
| gsl_sf_gamma_inc_Q_e | Y | gsl_sf_gamma_inc_Q(dbl_expr,dbl_expr)
|
| gsl_sf_gamma_inc_P_e | Y | gsl_sf_gamma_inc_P(dbl_expr,dbl_expr)
|
| gsl_sf_gamma_inc_e | Y | gsl_sf_gamma_inc(dbl_expr,dbl_expr)
|
| gsl_sf_lnbeta_e | Y | gsl_sf_lnbeta(dbl_expr,dbl_expr)
|
| gsl_sf_lnbeta_sgn_e | N | gsl_sf_lnbeta_sgn
|
| gsl_sf_beta_e | Y | gsl_sf_beta(dbl_expr,dbl_expr)
|
| gsl_sf_beta_inc_e | N | gsl_sf_beta_inc
|
| gsl_sf_gegenpoly_1_e | Y | gsl_sf_gegenpoly_1(dbl_expr,dbl_expr)
|
| gsl_sf_gegenpoly_2_e | Y | gsl_sf_gegenpoly_2(dbl_expr,dbl_expr)
|
| gsl_sf_gegenpoly_3_e | Y | gsl_sf_gegenpoly_3(dbl_expr,dbl_expr)
|
| gsl_sf_gegenpoly_n_e | N | gsl_sf_gegenpoly_n
|
| gsl_sf_gegenpoly_array | Y | gsl_sf_gegenpoly_array
|
| gsl_sf_hyperg_0F1_e | Y | gsl_sf_hyperg_0F1(dbl_expr,dbl_expr)
|
| gsl_sf_hyperg_1F1_int_e | Y | gsl_sf_hyperg_1F1_int(int_expr,int_expr,dbl_expr)
|
| gsl_sf_hyperg_1F1_e | Y | gsl_sf_hyperg_1F1(dbl_expr,dbl_expr,dbl_expr)
|
| gsl_sf_hyperg_U_int_e | Y | gsl_sf_hyperg_U_int(int_expr,int_expr,dbl_expr)
|
| gsl_sf_hyperg_U_int_e10_e | N | gsl_sf_hyperg_U_int_e10
|
| gsl_sf_hyperg_U_e | Y | gsl_sf_hyperg_U(dbl_expr,dbl_expr,dbl_expr)
|
| gsl_sf_hyperg_U_e10_e | N | gsl_sf_hyperg_U_e10
|
| gsl_sf_hyperg_2F1_e | Y | gsl_sf_hyperg_2F1(dbl_expr,dbl_expr,dbl_expr,dbl_expr)
|
| gsl_sf_hyperg_2F1_conj_e | Y | gsl_sf_hyperg_2F1_conj(dbl_expr,dbl_expr,dbl_expr,dbl_expr)
|
| gsl_sf_hyperg_2F1_renorm_e | Y | gsl_sf_hyperg_2F1_renorm(dbl_expr,dbl_expr,dbl_expr,dbl_expr)
|
| gsl_sf_hyperg_2F1_conj_renorm_e | Y | gsl_sf_hyperg_2F1_conj_renorm(dbl_expr,dbl_expr,dbl_expr,dbl_expr)
|
| gsl_sf_hyperg_2F0_e | Y | gsl_sf_hyperg_2F0(dbl_expr,dbl_expr,dbl_expr)
|
| gsl_sf_laguerre_1_e | Y | gsl_sf_laguerre_1(dbl_expr,dbl_expr)
|
| gsl_sf_laguerre_2_e | Y | gsl_sf_laguerre_2(dbl_expr,dbl_expr)
|
| gsl_sf_laguerre_3_e | Y | gsl_sf_laguerre_3(dbl_expr,dbl_expr)
|
| gsl_sf_laguerre_n_e | Y | gsl_sf_laguerre_n(int_expr,dbl_expr,dbl_expr)
|
| gsl_sf_lambert_W0_e | Y | gsl_sf_lambert_W0(dbl_expr)
|
| gsl_sf_lambert_Wm1_e | Y | gsl_sf_lambert_Wm1(dbl_expr)
|
| gsl_sf_legendre_Pl_e | Y | gsl_sf_legendre_Pl(int_expr,dbl_expr)
|
| gsl_sf_legendre_Pl_array | Y | status=gsl_sf_legendre_Pl_array(int,double,&var_out)
|
| gsl_sf_legendre_Pl_deriv_array | N | gsl_sf_legendre_Pl_deriv_array
|
| gsl_sf_legendre_P1_e | Y | gsl_sf_legendre_P1(dbl_expr)
|
| gsl_sf_legendre_P2_e | Y | gsl_sf_legendre_P2(dbl_expr)
|
| gsl_sf_legendre_P3_e | Y | gsl_sf_legendre_P3(dbl_expr)
|
| gsl_sf_legendre_Q0_e | Y | gsl_sf_legendre_Q0(dbl_expr)
|
| gsl_sf_legendre_Q1_e | Y | gsl_sf_legendre_Q1(dbl_expr)
|
| gsl_sf_legendre_Ql_e | Y | gsl_sf_legendre_Ql(int_expr,dbl_expr)
|
| gsl_sf_legendre_Plm_e | Y | gsl_sf_legendre_Plm(int_expr,int_expr,dbl_expr)
|
| gsl_sf_legendre_Plm_array | Y | status=gsl_sf_legendre_Plm_array(int,int,double,&var_out)
|
| gsl_sf_legendre_Plm_deriv_array | N | gsl_sf_legendre_Plm_deriv_array
|
| gsl_sf_legendre_sphPlm_e | Y | gsl_sf_legendre_sphPlm(int_expr,int_expr,dbl_expr)
|
| gsl_sf_legendre_sphPlm_array | Y | status=gsl_sf_legendre_sphPlm_array(int,int,double,&var_out)
|
| gsl_sf_legendre_sphPlm_deriv_array | N | gsl_sf_legendre_sphPlm_deriv_array
|
| gsl_sf_legendre_array_size | N | gsl_sf_legendre_array_size
|
| gsl_sf_conicalP_half_e | Y | gsl_sf_conicalP_half(dbl_expr,dbl_expr)
|
| gsl_sf_conicalP_mhalf_e | Y | gsl_sf_conicalP_mhalf(dbl_expr,dbl_expr)
|
| gsl_sf_conicalP_0_e | Y | gsl_sf_conicalP_0(dbl_expr,dbl_expr)
|
| gsl_sf_conicalP_1_e | Y | gsl_sf_conicalP_1(dbl_expr,dbl_expr)
|
| gsl_sf_conicalP_sph_reg_e | Y | gsl_sf_conicalP_sph_reg(int_expr,dbl_expr,dbl_expr)
|
| gsl_sf_conicalP_cyl_reg_e | Y | gsl_sf_conicalP_cyl_reg(int_expr,dbl_expr,dbl_expr)
|
| gsl_sf_legendre_H3d_0_e | Y | gsl_sf_legendre_H3d_0(dbl_expr,dbl_expr)
|
| gsl_sf_legendre_H3d_1_e | Y | gsl_sf_legendre_H3d_1(dbl_expr,dbl_expr)
|
| gsl_sf_legendre_H3d_e | Y | gsl_sf_legendre_H3d(int_expr,dbl_expr,dbl_expr)
|
| gsl_sf_legendre_H3d_array | N | gsl_sf_legendre_H3d_array
|
| gsl_sf_legendre_array_size | N | gsl_sf_legendre_array_size
|
| gsl_sf_log_e | Y | gsl_sf_log(dbl_expr)
|
| gsl_sf_log_abs_e | Y | gsl_sf_log_abs(dbl_expr)
|
| gsl_sf_complex_log_e | N | gsl_sf_complex_log
|
| gsl_sf_log_1plusx_e | Y | gsl_sf_log_1plusx(dbl_expr)
|
| gsl_sf_log_1plusx_mx_e | Y | gsl_sf_log_1plusx_mx(dbl_expr)
|
| gsl_sf_mathieu_a_array | N | gsl_sf_mathieu_a_array
|
| gsl_sf_mathieu_b_array | N | gsl_sf_mathieu_b_array
|
| gsl_sf_mathieu_a | N | gsl_sf_mathieu_a
|
| gsl_sf_mathieu_b | N | gsl_sf_mathieu_b
|
| gsl_sf_mathieu_a_coeff | N | gsl_sf_mathieu_a_coeff
|
| gsl_sf_mathieu_b_coeff | N | gsl_sf_mathieu_b_coeff
|
| gsl_sf_mathieu_ce | N | gsl_sf_mathieu_ce
|
| gsl_sf_mathieu_se | N | gsl_sf_mathieu_se
|
| gsl_sf_mathieu_ce_array | N | gsl_sf_mathieu_ce_array
|
| gsl_sf_mathieu_se_array | N | gsl_sf_mathieu_se_array
|
| gsl_sf_mathieu_Mc | N | gsl_sf_mathieu_Mc
|
| gsl_sf_mathieu_Ms | N | gsl_sf_mathieu_Ms
|
| gsl_sf_mathieu_Mc_array | N | gsl_sf_mathieu_Mc_array
|
| gsl_sf_mathieu_Ms_array | N | gsl_sf_mathieu_Ms_array
|
| gsl_sf_pow_int_e | N | gsl_sf_pow_int
|
| gsl_sf_psi_int_e | Y | gsl_sf_psi_int(int_expr)
|
| gsl_sf_psi_e | Y | gsl_sf_psi(dbl_expr)
|
| gsl_sf_psi_1piy_e | Y | gsl_sf_psi_1piy(dbl_expr)
|
| gsl_sf_complex_psi_e | N | gsl_sf_complex_psi
|
| gsl_sf_psi_1_int_e | Y | gsl_sf_psi_1_int(int_expr)
|
| gsl_sf_psi_1_e | Y | gsl_sf_psi_1(dbl_expr)
|
| gsl_sf_psi_n_e | Y | gsl_sf_psi_n(int_expr,dbl_expr)
|
| gsl_sf_synchrotron_1_e | Y | gsl_sf_synchrotron_1(dbl_expr)
|
| gsl_sf_synchrotron_2_e | Y | gsl_sf_synchrotron_2(dbl_expr)
|
| gsl_sf_transport_2_e | Y | gsl_sf_transport_2(dbl_expr)
|
| gsl_sf_transport_3_e | Y | gsl_sf_transport_3(dbl_expr)
|
| gsl_sf_transport_4_e | Y | gsl_sf_transport_4(dbl_expr)
|
| gsl_sf_transport_5_e | Y | gsl_sf_transport_5(dbl_expr)
|
| gsl_sf_sin_e | N | gsl_sf_sin
|
| gsl_sf_cos_e | N | gsl_sf_cos
|
| gsl_sf_hypot_e | N | gsl_sf_hypot
|
| gsl_sf_complex_sin_e | N | gsl_sf_complex_sin
|
| gsl_sf_complex_cos_e | N | gsl_sf_complex_cos
|
| gsl_sf_complex_logsin_e | N | gsl_sf_complex_logsin
|
| gsl_sf_sinc_e | N | gsl_sf_sinc
|
| gsl_sf_lnsinh_e | N | gsl_sf_lnsinh
|
| gsl_sf_lncosh_e | N | gsl_sf_lncosh
|
| gsl_sf_polar_to_rect | N | gsl_sf_polar_to_rect
|
| gsl_sf_rect_to_polar | N | gsl_sf_rect_to_polar
|
| gsl_sf_sin_err_e | N | gsl_sf_sin_err
|
| gsl_sf_cos_err_e | N | gsl_sf_cos_err
|
| gsl_sf_angle_restrict_symm_e | N | gsl_sf_angle_restrict_symm
|
| gsl_sf_angle_restrict_pos_e | N | gsl_sf_angle_restrict_pos
|
| gsl_sf_angle_restrict_symm_err_e | N | gsl_sf_angle_restrict_symm_err
|
| gsl_sf_angle_restrict_pos_err_e | N | gsl_sf_angle_restrict_pos_err
|
| gsl_sf_zeta_int_e | Y | gsl_sf_zeta_int(int_expr)
|
| gsl_sf_zeta_e | Y | gsl_sf_zeta(dbl_expr)
|
| gsl_sf_zetam1_e | Y | gsl_sf_zetam1(dbl_expr)
|
| gsl_sf_zetam1_int_e | Y | gsl_sf_zetam1_int(int_expr)
|
| gsl_sf_hzeta_e | Y | gsl_sf_hzeta(dbl_expr,dbl_expr)
|
| gsl_sf_eta_int_e | Y | gsl_sf_eta_int(int_expr)
|
| gsl_sf_eta_e | Y | gsl_sf_eta(dbl_expr)
|
As of version 3.9.9 (released July, 2009), NCO has wrappers to the GSL interpolation functions.
Given a set of data points (x1,y1)...(xn, yn) the GSL functions computes a continuous interpolating function Y(x) such that Y(xi) = yi. The interpolation is piecewise smooth, and its behavior at the end-points is determined by the type of interpolation used. For more information consult the GSL manual.
Interpolation with ncap2 is a two stage process. In the first stage, a ram variable is created from the chosen interpolating function and the data set. This ram variable holds in memory a GSL interpolation object. In the second stage, points along the interpolating function are calculated. If you have a very large data set or are interpolating many sets then consider deleting the ram variable when it is redundant. Use the command ram_delete(var_nm).
A simple example
x_in[$lon]={1.0,2.0,3.0,4.0};
y_in[$lon]={1.1,1.2,1.5,1.8};
// Ram variable is declared and defined here
gsl_interp_cspline(&ram_sp,x_in,y_in);
x_out[$lon_grd]={1.1,2.0,3.0,3.1,3.99};
y_out=gsl_spline_eval(ram_sp,x_out);
y2=gsl_spline_eval(ram_sp,1.3);
y3=gsl_spline_eval(ram_sp,0.0);
ram_delete(ram_sp);
print(y_out); // 1.10472, 1.2, 1.4, 1.42658, 1.69680002
print(y2); // 1.12454
print(y3); // '_'
Note in the above example y3 is set to 'missing value' because 0.0 isn't within the input X range.
GSL Interpolation Types
All the interpolation functions have been implemented. These are:
gsl_interp_linear()
gsl_interp_polynomial()
gsl_interp_cspline()
gsl_interp_cspline_periodic()
gsl_interp_akima()
gsl_interp_akima_periodic()
Evaluation of Interpolating Types
Implemented
gsl_spline_eval()
Unimplemented
gsl_spline_deriv()
gsl_spline_deriv2()
gsl_spline_integ()
See the ncap.in and ncap2.in scripts released with NCO for more complete demonstrations of ncap and ncap2 functionality, respectively (these scripts are available on-line at http://nco.sf.net/ncap.in and http://nco.sf.net/ncap2.in).
Define new attribute new for existing variable one as twice the existing attribute double_att of variable att_var:
ncap2 -s 'one@new=2*att_var@double_att' in.nc out.nc
Average variables of mixed types (result is of type double):
ncap2 -s 'average=(var_float+var_double+var_int)/3' in.nc out.nc
Multiple commands may be given to ncap2 in three ways.
First, the commands may be placed in a script which is executed, e.g.,
tst.nco.
Second, the commands may be individually specified with multiple
‘-s’ arguments to the same ncap2 invocation.
Third, the commands may be chained together into a single ‘-s’
argument to ncap2.
Assuming the file tst.nco contains the commands
a=3;b=4;c=sqrt(a^2+b^2);, then the following ncap2
invocations produce identical results:
ncap2 -v -S tst.nco in.nc out.nc
ncap2 -v -s 'a=3' -s 'b=4' -s 'c=sqrt(a^2+b^2)' in.nc out.nc
ncap2 -v -s 'a=3;b=4;c=sqrt(a^2+b^2)' in.nc out.nc
The second and third examples show that ncap2 does not require that a trailing semi-colon ‘;’ be placed at the end of a ‘-s’ argument, although a trailing semi-colon ‘;’ is always allowed. However, semi-colons are required to separate individual assignment statements chained together as a single ‘-s’ argument.
ncap2 may be used to “grow” dimensions, i.e., to increase
dimension sizes without altering existing data.
Say in.nc has ORO(lat,lon) and the user wishes a new
file with new_ORO(new_lat,new_lon) that contains zeros in the
undefined portions of the new grid.
defdim("new_lat",$lat.size+1); // Define new dimension sizes
defdim("new_lon",$lon.size+1);
new_ORO[$new_lat,$new_lon]=0.0f; // Initialize to zero
new_ORO(0:$lat.size-1,0:$lon.size-1)=ORO; // Fill valid data
The commands to define new coordinate variables new_lat
and new_lon in the output file follow a similar pattern.
One would might store these commands in a script grow.nco
and then execute the script with
ncap2 -v -S grow.nco in.nc out.nc
Imagine you wish to create a binary flag based on the value of
an array.
The flag should have value 1.0 where the array exceeds 1.0,
and value 0.0 elsewhere.
This example creates the binary flag ORO_flg in out.nc
from the continuous array named ORO in in.nc.
ncap2 -s 'ORO_flg=(ORO > 1.0)' in.nc out.nc
Suppose your task is to change all values of ORO which
equal 2.0 to the new value 3.0:
ncap2 -s 'ORO_msk=(ORO==2.0);ORO=ORO_msk*3.0+!ORO_msk*ORO' in.nc out.nc
This creates and uses ORO_msk to mask the subsequent arithmetic
operation.
Values of ORO are only changed where ORO_msk is true,
i.e., where ORO equals 2.0
Using the where statement the above code simplifies to :
ncap2 -s 'where(ORO==2.0) ORO=3.0;' in.nc foo.nc
This example uses ncap2 to compute the covariance of two variables. Let the variables u and v be the horizontal wind components. The covariance of u and v is defined as the time mean product of the deviations of u and v from their respective time means. Symbolically, the covariance
[u'v'] = [uv]-[u][v] where [x] denotes the time-average of x and x'
denotes the deviation from the time-mean.
The covariance tells us how much of the correlation of two signals
arises from the signal fluctuations versus the mean signals.
Sometimes this is called the eddy covariance.
We will store the covariance in the variable uprmvprm.
ncwa -O -a time -v u,v in.nc foo.nc # Compute time mean of u,v
ncrename -O -v u,uavg -v v,vavg foo.nc # Rename to avoid conflict
ncks -A -v uavg,vavg foo.nc in.nc # Place time means with originals
ncap2 -O -s 'uprmvprm=u*v-uavg*vavg' in.nc in.nc # Covariance
ncra -O -v uprmvprm in.nc foo.nc # Time-mean covariance
The mathematically inclined will note that the same covariance would be obtained by replacing the step involving ncap2 with
ncap2 -O -s 'uprmvprm=(u-uavg)*(v-vavg)' foo.nc foo.nc # Covariance
As of NCO version 3.1.8 (December, 2006), ncap2 can compute averages, and thus covariances, by itself:
ncap2 -s 'uavg=u.avg($time);vavg=v.avg($time);uprmvprm=u*v-uavg*vavg' \
-s 'uprmvrpmavg=uprmvprm.avg($time)' in.nc foo.nc
We have not seen a simpler method to script and execute powerful arithmetic than ncap2.
ncap2 utilizes many meta-characters
(e.g., ‘$’, ‘?’, ‘;’, ‘()’, ‘[]’)
that can confuse the command-line shell if not quoted properly.
The issues are the same as those which arise in utilizing extended
regular expressions to subset variables (see Subsetting Variables).
The example above will fail with no quotes and with double quotes.
This is because shell globbing tries to interpolate the value of
$time from the shell environment unless it is quoted:
ncap2 -s 'uavg=u.avg($time)' in.nc foo.nc # Correct (recommended)
ncap2 -s uavg=u.avg('$time') in.nc foo.nc # Correct (and dangerous)
ncap2 -s uavg=u.avg($time) in.nc foo.nc # Fails ($time = '')
ncap2 -s "uavg=u.avg($time)" in.nc foo.nc # Fails ($time = '')
Without the single quotes, the shell replaces $time with an
empty string.
The command ncap2 receives from the shell is
uavg=u.avg().
This causes ncap2 to average over all dimensions rather than
just the time dimension, and unintended consequence.
We recommend using single quotes to protect ncap2 command-line scripts from the shell, even when such protection is not strictly necessary. Expert users may violate this rule to exploit the ability to use shell variables in ncap2 command-line scripts (see CCSM Example). In such cases it may be necessary to use the shell backslash character ‘\’ to protect the ncap2 meta-character.
Whether a degenerate record dimension is desirable or undesirable
depends on the application.
Often a degenerate time dimension is useful, e.g., for
concatentating, but it may cause problems with arithmetic.
Such is the case in the above example, where the first step employs
ncwa rather than ncra for the time-averaging.
Of course the numerical results are the same with both operators.
The difference is that, unless ‘-b’ is specified, ncwa
writes no time dimension to the output file, while ncra
defaults to keeping time as a degenerate (size 1) dimension.
Appending u and v to the output file would cause
ncks to try to expand the degenerate time axis of uavg
and vavg to the size of the non-degenerate time dimension
in the input file.
Thus the append (ncks -A) command would be undefined (and
should fail) in this case.
Equally important is the ‘-C’ argument
(see Subsetting Coordinate Variables) to ncwa to prevent
any scalar time variable from being written to the output file.
Knowing when to use ncwa -a time rather than the default
ncra for time-averaging takes, well, time.
ncap2 supports the standard mathematical functions supplied with most operating systems. Standard calculator notation is used for addition +, subtraction -, multiplication *, division /, exponentiation ^, and modulus %. The available elementary mathematical functions are:
abs(x)acos(x)acosh(x)asin(x)asinh(x)atan(x)atan2(y,x)atanh(x)ceil(x)cos(x)cosh(x)erf(x)erfc(x)exp(x)floor(x)gamma(x)gamma_inc_P(x)ln(x)log(x)ln(x).
log10(x)nearbyint(x)pow(x,y)pow function,
integer arguments are promoted (see Type Conversion) to type
NC_FLOAT before evaluation.
Example:
pow(2,3) = 8
rint(x)round(x)sin(x)sinh(x)sqrt(x)tan(x)tanh(x)trunc(x)This page lists the ncap operators in order of precedence (highest to lowest). Their associativity indicates in what order operators of equal precedence in an expression are applied.
| Operator | Description | Associativity
|
|---|---|---|
++ -- | Postfix Increment/Decrement | Right to Left
|
() | Parentheses (function call)
| |
. | Method call
| |
++ -- | Prefix Increment/Decrement | Right to Left
|
+ - | Unary Plus/Minus
| |
! | Logical Not
| |
^ | Power of Operator | Right to Left
|
* / % | Multiply/Divide/Modulus | Left To Right
|
+ - | Addition/Subtraction | Left To Right
|
>> << | Fortran style array clipping | Left to Right
|
< <= | Less than/Less than or equal to | Left to Right
|
> >= | Greater than/Greater than or equal to
| |
== != | Equal to/Not equal to | Left to Right
|
&& | Logical AND | Left to Right
|
|| | Logical OR | Left to Right
|
?: | Ternary Operator | Right to Left
|
= | Assignment | Right to Left
|
+= -= | Addition/subtraction assignment
| |
*= /= | Multiplication/division assignment
|
In this section when I refer to a name I mean a variable name, attribute name or a dimension name The allowed characters in a valid netCDF name vary from release to release. (See end section). If you want to use metacharacters in a name or use a method name as a variable name then the name has to be quoted wherever it occurs.
The default nco name is specified by the regular expressions:
DGT: ('0'..'9');
LPH: ( 'a'..'z' | 'A'..'Z' | '_' );
name: (LPH)(LPH|DGT)+
The first character of a valid name must be alphabetic or the underscore. Any subsequent characters must be alphanumeric or underscore. ( e.g a1,_23, hell_is_666 )
The valid characters in a quoted name are specified by the regular expressions:
LPHDGT: ( 'a'..'z' | 'A'..'Z' | '_' | '0'..'9');
name: (LPHDGT|'-'|'+'|'.'|'('|')'|':' )+ ;
Quote a variable:
'avg' , '10_+10','set_miss' '+-90field' , '–test'=10.0d
Quote a attribute:
'three@10', 'set_mss@+10', '666@hell', 't1@+units'="kelvin"
Quote a dimension:
'$10', '$t1–', '$–odd', c1['$10','$t1–']=23.0d
The following comments are lifted directly from the netcdf libraries and detail the naming conventions for each release.
netcdf-3.5.1
netcdf-3.6.0-p1
netcdf-3.6.1
netcdf-3.6.2
/*
* ( [a-zA-Z]|[0-9]|'_'|'-'|'+'|'.'|'|':'|'@'|'('|')' )+
* Verify that a name string is valid
* CDL syntax, eg, all the characters are
* alphanumeric, '-', '_', '+', or '.'.
* Also permit ':', '@', '(', or ')' in names for chemists currently making
* use of these characters, but don't document until ncgen and ncdump can
* also handle these characters in names.
*/
netcdf-3.6.3
netcdf-4.0 Final 2008/08/28
/*
* Verify that a name string is valid syntax. The allowed name
* syntax (in RE form) is:
*
* ([a-zA-Z_]|{UTF8})([^\x00-\x1F\x7F/]|{UTF8})*
*
* where UTF8 represents a multibyte UTF-8 encoding. Also, no
* trailing spaces are permitted in names. This definition
* must be consistent with the one in ncgen.l. We do not allow '/'
* because HDF5 does not permit slashes in names as slash is used as a
* group separator. If UTF-8 is supported, then a multi-byte UTF-8
* character can occur anywhere within an identifier. We later
* normalize UTF-8 strings to NFC to facilitate matching and queries.
*/
ncatted [-a att_dsc] [-a ...] [-D dbg] [-h] [--hdr_pad nbr]
[-l path] [-O] [-o output-file] [-p path] [-R] [-r]
input-file [[output-file]]
DESCRIPTION
ncatted edits attributes in a netCDF file.
If you are editing attributes then you are spending too much time in the
world of metadata, and ncatted was written to get you back out as
quickly and painlessly as possible.
ncatted can append, create, delete,
modify, and overwrite attributes (all explained below).
Furthermore, ncatted allows each editing operation to be applied
to every variable in a file.
This saves time when changing attribute conventions throughout a file.
Note that ncatted interprets character attributes
(i.e., attributes of type NC_CHAR) as strings.
Because repeated use of ncatted can considerably increase the size
of the history global attribute (see History Attribute), the
‘-h’ switch is provided to override automatically appending the
command to the history global attribute in the output-file.
When ncatted is used to change the _FillValue attribute,
it changes the associated missing data self-consistently.
If the internal floating point representation of a missing value,
e.g., 1.0e36, differs between two machines then netCDF files produced
on those machines will have incompatible missing values.
This allows ncatted to change the missing values in files from
different machines to a single value so that the files may then be
concatenated together, e.g., by ncrcat, without losing any
information.
See Missing Values, for more information.
The key to mastering ncatted is understanding the meaning of the
structure describing the attribute modification, att_dsc specified by the required option ‘-a’ or ‘--attribute’.
Each att_dsc contains five elements, which makes using
ncatted somewhat complicated, but powerful.
The att_dsc argument structure contains five arguments in the
following order:
att_dsc = att_nm, var_nm, mode, att_type,
att_val
units
pressure, '^H2O'.
a.
See below for complete listing of valid values of mode.
c.
See below for complete listing of valid values of att_type.
pascal.
The value of att_nm is the name of the attribute you want to edit. This meaning of this should be clear to all users of the ncatted operator. If att_nm is omitted (i.e., left blank) and Delete mode is selected, then all attributes associated with the specified variable will be deleted.
The value of var_nm is the name of the variable containing the attribute (named att_nm) that you want to edit. There are three very important and useful exceptions to this rule. The value of var_nm can also be used to direct ncatted to edit global attributes, or to repeat the editing operation for every variable in a file. A value of var_nm of “global” indicates that att_nm refers to a global attribute, rather than a particular variable's attribute. This is the method ncatted supports for editing global attributes. If var_nm is left blank, on the other hand, then ncatted attempts to perform the editing operation on every variable in the file. This option may be convenient to use if you decide to change the conventions you use for describing the data. Finally, as mentioned above, var_nm may be specified as a regular expression.
The value of mode is a single character abbreviation (a,
c, d, m, or o) standing for one of
five editing modes:
acdmo The value of att_type is a single character abbreviation
(f, d, l, i, s, c,
b, u) or a short string standing for one of the twelve
primitive netCDF data types:
fNC_FLOAT.
dNC_DOUBLE.
i, lNC_INT.
sNC_SHORT.
cNC_CHAR.
bNC_BYTE.
ubNC_UBYTE.
usNC_USHORT.
u, ui, ulNC_UINT.
ll, int64NC_INT64.
ull, uint64NC_UINT64.
sngNC_STRING.
The value of att_val is what you want to change attribute
att_nm to contain.
The specification of att_val is optional in Delete (and is
ignored) mode.
Attribute values for all types besides NC_CHAR must have an
attribute length of at least one.
Thus att_val may be a single value or one-dimensional array of
elements of type att_type.
If the att_val is not set or is set to empty space,
and the att_type is NC_CHAR, e.g., -a units,T,o,c,""
or -a units,T,o,c,, then the corresponding attribute is set to
have zero length.
When specifying an array of values, it is safest to enclose
att_val in single or double quotes, e.g.,
-a levels,T,o,s,"1,2,3,4" or
-a levels,T,o,s,'1,2,3,4'.
The quotes are strictly unnecessary around att_val except
when att_val contains characters which would confuse the calling
shell, such as spaces, commas, and wildcard characters.
NCO processing of NC_CHAR attributes is a bit like Perl in
that it attempts to do what you want by default (but this sometimes
causes unexpected results if you want unusual data storage).
If the att_type is NC_CHAR then the argument is interpreted as a
string and it may contain C-language escape sequences, e.g., \n,
which NCO will interpret before writing anything to disk.
NCO translates valid escape sequences and stores the
appropriate ASCII code instead.
Since two byte escape sequences, e.g., \n, represent one-byte
ASCII codes, e.g., ASCII 10 (decimal), the stored
string attribute is one byte shorter than the input string length for
each embedded escape sequence.
The most frequently used C-language escape sequences are \n (for
linefeed) and \t (for horizontal tab).
These sequences in particular allow convenient editing of formatted text
attributes.
The other valid ASCII codes are \a, \b, \f,
\r, \v, and \\.
See ncks netCDF Kitchen Sink, for more examples of string formatting
(with the ncks ‘-s’ option) with special characters.
Analogous to printf, other special characters are also allowed by
ncatted if they are "protected" by a backslash.
The characters ", ', ?, and \ may be
input to the shell as \", \', \?, and \\.
NCO simply strips away the leading backslash from these
characters before editing the attribute.
No other characters require protection by a backslash.
Backslashes which precede any other character (e.g., 3, m,
$, |, &, @, %, {, and
}) will not be filtered and will be included in the attribute.
Note that the NUL character \0 which terminates C language
strings is assumed and need not be explicitly specified.
If \0 is input, it will not be translated (because it would
terminate the string in an additional location).
Because of these context-sensitive rules, if wish to use an attribute of
type NC_CHAR to store data, rather than text strings, you should use
ncatted with care.
Append the string "Data version 2.0.\n" to the global attribute
history:
ncatted -a history,global,a,c,"Data version 2.0\n" in.nc
Note the use of embedded C language printf()-style escape
sequences.
Change the value of the long_name attribute for variable T
from whatever it currently is to "temperature":
ncatted -a long_name,T,o,c,temperature in.nc
Delete all existing units attributes:
ncatted -a units,,d,, in.nc
The value of var_nm was left blank in order to select all variables in the file. The values of att_type and att_val were left blank because they are superfluous in Delete mode.
Delete all attributes associated with the tpt variable:
ncatted -a ,tpt,d,, in.nc
The value of att_nm was left blank in order to select all
attributes associated with the variable.
To delete all global attributes, simply replace tpt with
global in the above.
Modify all existing units attributes to "meter second-1":
ncatted -a units,,m,c,"meter second-1" in.nc
Add a units attribute of "kilogram kilogram-1" to all variables
whose first three characters are ‘H2O’:
ncatted -a units,'^H2O',c,c,"kilogram kilogram-1" in.nc
Overwrite the quanta attribute of variable
energy to an array of four integers.
ncatted -O -a quanta,energy,o,s,"010,101,111,121" in.nc
As of NCO 3.9.6 (January, 2009), variable names arguments
to ncatted may contain extended regular expressions.
Create isotope attributes for all variables containing ‘H2O’
in their names.
ncatted -O -a isotope,'^H2O*',c,s,"18" in.nc
See Subsetting Variables for more details.
Demonstrate input of C-language escape sequences (e.g., \n) and
other special characters (e.g., \")
ncatted -h -a special,global,o,c,
'\nDouble quote: \"\nTwo consecutive double quotes: \"\"\n
Single quote: Beyond my shell abilities!\nBackslash: \\\n
Two consecutive backslashes: \\\\\nQuestion mark: \?\n' in.nc
Note that the entire attribute is protected from the shell by single quotes. These outer single quotes are necessary for interactive use, but may be omitted in batch scripts.
ncbo [-3] [-4] [-6] [-A] [-C] [-c] [-D dbg]
[-d dim,[min][,[max][,[stride]]] [-F] [-h]
[-L dfl_lvl] [-l path] [-O] [-o file_3] [-p path] [-R] [-r]
[-t thr_nbr] [-v var[,...]] [-X ...] [-x] [-y op_typ]
file_1 file_2 [file_3]
DESCRIPTION
ncbo performs binary operations on variables in file_1 and the corresponding variables (those with the same name) in file_2 and stores the results in file_3. The binary operation operates on the entire files (modulo any excluded variables). See Missing Values, for treatment of missing values. One of the four standard arithmetic binary operations currently supported must be selected with the ‘-y op_typ’ switch (or long options ‘--op_typ’ or ‘--operation’). The valid binary operations for ncbo, their definitions, corresponding values of the op_typ key, and alternate invocations are:
ncbo --op_typ=* 1.nc 2.nc 3.nc # Dangerous (shell may try to glob)
ncbo --op_typ='*' 1.nc 2.nc 3.nc # Safe ('*' protected from shell)
ncbo --op_typ="*" 1.nc 2.nc 3.nc # Safe ('*' protected from shell)
ncbo --op_typ=mlt 1.nc 2.nc 3.nc
ncbo --op_typ=mult 1.nc 2.nc 3.nc
ncbo --op_typ=multiply 1.nc 2.nc 3.nc
ncbo --op_typ=multiplication 1.nc 2.nc 3.nc
ncmult 1.nc 2.nc 3.nc # First do 'ln -s ncbo ncmult'
ncmultiply 1.nc 2.nc 3.nc # First do 'ln -s ncbo ncmultiply'
No particular argument or invocation form is preferred. Users are encouraged to use the forms which are most intuitive to them.
Normally, ncbo will fail unless an operation type is specified with ‘-y’ (equivalent to ‘--op_typ’). You may create exceptions to this rule to suit your particular tastes, in conformance with your site's policy on symbolic links to executables (files of a different name point to the actual executable). For many years, ncdiff was the main binary file operator. As a result, many users prefer to continue invoking ncdiff rather than memorizing a new command (‘ncbo -y sbt’) which behaves identically to the original ncdiff command. However, from a software maintenance standpoint, maintaining a distinct executable for each binary operation (e.g., ncadd) is untenable, and a single executable, ncbo, is desirable. To maintain backward compatibility, therefore, NCO automatically creates a symbolic link from ncbo to ncdiff. Thus ncdiff is called an alternate invocation of ncbo. ncbo supports many additional alternate invocations which must be manually activated. Should users or system adminitrators decide to activate them, the procedure is simple. For example, to use ‘ncadd’ instead of ‘ncbo --op_typ=add’, simply create a symbolic link from ncbo to ncadd 37. The alternatate invocations supported for each operation type are listed above. Alternatively, users may always define ‘ncadd’ as an alias to ‘ncbo --op_typ=add’ 38.
It is important to maintain portability in NCO scripts. Therefore we recommend that site-specfic invocations (e.g., ‘ncadd’) be used only in interactive sessions from the command-line. For scripts, we recommend using the full invocation (e.g., ‘ncbo --op_typ=add’). This ensures portability of scripts between users and sites.
ncbo operates (e.g., adds) variables in file_2 with the
corresponding variables (those with the same name) in file_1 and
stores the results in file_3.
Variables in file_2 are broadcast to conform to the
corresponding variable in file_1 if necessary, but the reverse is
not true.
Broadcasting a variable means creating data in non-existing dimensions
from the data in existing dimensions.
For example, a two dimensional variable in file_2 can be
subtracted from a four, three, or two (but not one or zero)
dimensional variable (of the same name) in file_1.
This functionality allows the user to compute anomalies from the mean.
Note that variables in file_1 are not broadcast to conform
to the dimensions in file_2.
In the future, we will broadcast variables in file_1, if necessary
to conform to their counterparts in file_2.
Thus, presently, the number of dimensions, or rank, of any
processed variable in file_1 must be greater than or equal to the
rank of the same variable in file_2.
Furthermore, the size of all dimensions common to both file_1 and
file_2 must be equal.
When computing anomalies from the mean it is often the case that
file_2 was created by applying an averaging operator to a file
with initially the same dimensions as file_1 (often file_1
itself).
In these cases, creating file_2 with ncra rather than
ncwa will cause the ncbo operation to fail.
For concreteness say the record dimension in file_1 is
time.
If file_2 were created by averaging file_1 over the
time dimension with the ncra operator rather than with
the ncwa operator, then file_2 will have a time
dimension of size 1 rather than having no time dimension at
all
39.
In this case the input files to ncbo, file_1 and
file_2, will have unequally sized time dimensions which
causes ncbo to fail.
To prevent this from occuring, use ncwa to remove the
time dimension from file_2.
See the example below.
ncbo never operates on coordinate variables or variables
of type NC_CHAR or NC_BYTE.
This ensures that coordinates like (e.g., latitude and longitude) are
physically meaningful in the output file, file_3.
This behavior is hardcoded.
ncbo applies special rules to some
CF-defined (and/or NCAR CCSM or NCAR CCM
fields) such as ORO.
See CF Conventions for a complete description.
Finally, we note that ncflint (see ncflint netCDF File Interpolator) is designed for file interpolation.
As such, it also performs file subtraction, addition, multiplication,
albeit in a more convoluted way than ncbo.
Say files 85_0112.nc and 86_0112.nc each contain 12 months of data. Compute the change in the monthly averages from 1985 to 1986:
ncbo -op_typ=sub 86_0112.nc 85_0112.nc 86m85_0112.nc
ncdiff 86_0112.nc 85_0112.nc 86m85_0112.nc
The following examples demonstrate the broadcasting feature of
ncbo.
Say we wish to compute the monthly anomalies of T from the yearly
average of T for the year 1985.
First we create the 1985 average from the monthly data, which is stored
with the record dimension time.
ncra 85_0112.nc 85.nc
ncwa -O -a time 85.nc 85.nc
The second command, ncwa, gets rid of the time dimension
of size 1 that ncra left in 85.nc.
Now none of the variables in 85.nc has a time dimension.
A quicker way to accomplish this is to use ncwa from the
beginning:
ncwa -a time 85_0112.nc 85.nc
We are now ready to use ncbo to compute the anomalies for 1985:
ncdiff -v T 85_0112.nc 85.nc t_anm_85_0112.nc
Each of the 12 records in t_anm_85_0112.nc now contains the
monthly deviation of T from the annual mean of T for each
gridpoint.
Say we wish to compute the monthly gridpoint anomalies from the zonal
annual mean.
A zonal mean is a quantity that has been averaged over the
longitudinal (or x) direction.
First we use ncwa to average over longitudinal direction
lon, creating 85_x.nc, the zonal mean of 85.nc.
Then we use ncbo to subtract the zonal annual means from the
monthly gridpoint data:
ncwa -a lon 85.nc 85_x.nc
ncdiff 85_0112.nc 85_x.nc tx_anm_85_0112.nc
This examples works assuming 85_0112.nc has dimensions
time and lon, and that 85_x.nc has no time
or lon dimension.
As a final example, say we have five years of monthly data (i.e., 60 months) stored in 8501_8912.nc and we wish to create a file which contains the twelve month seasonal cycle of the average monthly anomaly from the five-year mean of this data. The following method is just one permutation of many which will accomplish the same result. First use ncwa to create the five-year mean:
ncwa -a time 8501_8912.nc 8589.nc
Next use ncbo to create a file containing the difference of each month's data from the five-year mean:
ncbo 8501_8912.nc 8589.nc t_anm_8501_8912.nc
Now use ncks to group the five January anomalies together in one file, and use ncra to create the average anomaly for all five Januarys. These commands are embedded in a shell loop so they are repeated for all twelve months:
for idx in {1..12}; do # Bash Shell (version 3.0+)
idx=`printf "%02d" ${idx}` # Zero-pad to preserve order
ncks -F -d time,${idx},,12 t_anm_8501_8912.nc foo.${idx}
ncra foo.${idx} t_anm_8589_${idx}.nc
done
for idx in 01 02 03 04 05 06 07 08 09 10 11 12; do # Bourne Shell
ncks -F -d time,${idx},,12 t_anm_8501_8912.nc foo.${idx}
ncra foo.${idx} t_anm_8589_${idx}.nc
done
foreach idx (01 02 03 04 05 06 07 08 09 10 11 12) # C Shell
ncks -F -d time,${idx},,12 t_anm_8501_8912.nc foo.${idx}
ncra foo.${idx} t_anm_8589_${idx}.nc
end
Note that ncra understands the stride argument so the
two commands inside the loop may be combined into the single command
ncra -F -d time,${idx},,12 t_anm_8501_8912.nc foo.${idx}
Finally, use ncrcat to concatenate the 12 average monthly anomaly files into one twelve-record file which contains the entire seasonal cycle of the monthly anomalies:
ncrcat t_anm_8589_??.nc t_anm_8589_0112.nc
ncea [-3] [-4] [-6] [-A] [-C] [-c] [-D dbg]
[-d dim,[min][,[max][,[stride]]] [-F] [-h] [-L dfl_lvl] [-l path]
[-n loop] [-O] [-o output-file] [-p path] [-R] [-r]
[-t thr_nbr] [-v var[,...]] [-X ...] [-x] [-y op_typ]
[input-files] [output-file]
DESCRIPTION
ncea performs gridpoint averages of variables across an arbitrary number (an ensemble) of input-files, with each file receiving an equal weight in the average. ncea averages entire files, and weights each file evenly. This is distinct from ncra, which only averages over the record dimension (e.g., time), and weights each record in the record dimension evenly,
Variables in the output-file are the same size as the variable in each of the input-files, and all input-files must be the same size. The only exception is that ncea allows files to differ in the record dimension size if the requested record hyperslab (see Hyperslabs) resolves to the same size for all files. ncea recomputes the record dimension hyperslab limits for each input file so that coordinate limits may be used to select equal length timeseries from unequal length files. This simplifies analysis of unequal lengt