CDM utility programs

Below are useful command-line utilites that can be called from the CDM library. The easiest way to use these is to grab the latest netcdfAll.jar file.

• ncdump: prints the textual representation of a dataset to standard output
• nccopy: copies a CDM dataset to a netCDF-3 (default) or netCDF-4 file
• nccompare: compares two CDM files for semantic equivilence

• BufrSplitter: copies a BUFR file's messages into separate output files, depending on type
• CFPointWriter: copies a CDM point feature dataset to CF/NetCDF format
• GribCdmIndex: write GRIB Collection Indexes
• FeatureScan: scans a directory to find CDM datasets and determines their FeatureTypes

ncdump

Prints the textual representation of a dataset to standard output. Similar functionality to the ncdump utility program. By default, just the header (ncdump -h) is output. This application works on any CDM file, not just netCDF files.

java -Xmx1g -classpath netcdfAll-<version>.jar ucar.nc2.NCdumpW
    filepath [-cdl | -ncml] [-c | -vall] [-v varName1;varName2;..] [-v varName(0:1,:,12)]

where:

• filepath: pathname of any CDM file.
• -cdl : show CDL (strict mode)
• -ncml: show NcML (default)
• -c: show data for coordinate variables only
• -vall: show data for all variables
• -v varName1;varName2;..: show data for these variables, use variable's full names (including groups if present)
• -v varName(0:1,:,12): show data for a section of this variable only, using F90 section specificiation

nccopy

Copies a dataset to a netCDF-3 (default) or netCDF-4 file. The input may be any CDM dataset, including OPeNDAP URLs, NcML, GRIB files, etc. If the dataset uses the extended data model, you must write to netCDF-4. If writing to netCDF-4, you must have the netcdf-4 C library loaded.

java -Xmx1g -classpath netcdfAll-<version>.jar ucar.nc2.write.Nccopy [options]
  Options:
  * -i, --input
       Input dataset.
  * -o, --output
       Output file.
    -f, --format
       Output file format. Allowed values = [netcdf3, netcdf4, netcdf4_classic,
       netcdf3c, netcdf3c64, ncstream]
       Default: netcdf3
    -isLargeFile, --isLargeFile
       Write to large file format. Only used in NetCDF 3. Allowed values =
       [standard, grib, none]
       Default: false
    -st, --strategy
       Chunking strategy. Only used in NetCDF 4. Allowed values = [standard,
       grib, none]
       Default: standard
    -d, --deflateLevel
       Compression level. Only used in NetCDF 4. Allowed values = 0 (no
       compression, fast) to 9 (max compression, slow)
       Default: 5
    -sh, --shuffle
       Enable the shuffle filter, which may improve compression. Only used in
       NetCDF 4. This option is ignored unless a non-zero deflate level is specified.
       Default: true
    -h, --help
       Display this help and exit
       Default: false

nccompare

Compares two CDM files for semantic equivilence.

java -Xmx1g -classpath netcdfAll-<version>.jar ucar.nc2.util.CompareNetcdf2
    file1 file2 [-showEach] [-compareData]

where

• -showEach: show details of comparing each object
• -compareData: compare data also
• file1: first file to compare
• file2: second file to compare

BufrSplitter

Copies a BUFR file's messages into separate output files, depending on message type.

java -Xmx1g -classpath netcdfAll-<version>.jar ucar.nc2.iosp.bufr.writer.BufrSplitter
    --fileSpec <fileIn> --dirOut <dirOut>

where

• --fileSpec: file to split
• --dirOut: output directory of split operation

CFPointWriter

Copies a CDM point feature dataset to CF/NetCDF format. The CF conventions target NetCDF-3, but you can also write NetCDF-4 files in classic mode. For that, you must first install the C library.

java -Xmx1g -classpath netcdfAll-<version>.jar ucar.nc2.ft.point.writer.CFPointWriter [options]
  Options:
  * -i, --input
       Input file.
  * -o, --output
       Output file.
    -f, --format
       Output file format. Allowed values = [netcdf3, netcdf4, netcdf4_classic,
       netcdf3c, netcdf3c64, ncstream]
       Default: netcdf3
    -st, --strategy
       Chunking strategy. Only used in NetCDF 4. Allowed values = [standard,
       grib, none]
       Default: standard
    -d, --deflateLevel
       Compression level. Only used in NetCDF 4. Allowed values = 0 (no
       compression, fast) to 9 (max compression, slow)
       Default: 5
    -sh, --shuffle
       Enable the shuffle filter, which may improve compression. Only used in
       NetCDF 4. This option is ignored unless a non-zero deflate level is specified.
       Default: true
    -h, --help
       Display this help and exit
       Default: false

GribCdmIndex

Write GRIB Collection Indexes from an XML file containing a GRIB <featureCollection> XML element.

java -Xmx1g -classpath netcdfAll-<version>.jar ucar.nc2.grib.collection.GribCdmIndex [options]
  Options:
  * -fc, --featureCollection
       Input XML file containing <featureCollection> root element
    -update, --CollectionUpdateType
       Collection Update Type
       Default: always
    -h, --help
       Display this help and exit
       Default: false

Example:

java -Xmx1g -classpath netcdfAll-<version>.jar ucar.nc2.grib.collection.GribCdmIndex -fc /data/fc/gfs_example.xml

Note that the output file is placed in the root directory of the collection, as specified by the Collection Specification of the GRIB <featureCollection>.

FeatureScan

Scans all the files in a directory to see if they are CDM files and can be identified as a particular feature type.

java -Xmx1g -classpath netcdfAll-<version>.jar ucar.nc2.ft.scan.FeatureScan directory [-subdirs]

where

• directory: scan this directory
• -subdirs: recurse into subdirectories

This document was last updated Mar 2015.