Project Management Service

• cdo option -z zip: added optional compression level -z zip[_1-9]
• cdo: added option -k <chunktype> to set the chunk type to auto, grid or lines
• Added workaround to combine CDO operators with the result of mergetime, merge, copy, cat, ens<STAT>
  (use one input parameter with wildcards in single quotes, e.g.: 'ifile?_*')

• enlarge: added missing value support

• gradsdes: failed
• sellevel: loosing level bounds
• wrong result for user defined lonlat grids with xfirst < 0
  (This bug was introduced in CDO version 1.5.8)

• file.c::file_initialize: added support for env GRIB_API_IO_BUFFER_SIZE

• vlistDefVarChunktype(): set chunktype to CHUNK_AUTO, CHUNK_GRID, CHUNK_LINES
• vlistInqVarChunktype(): returns the chunktype

• vlistCopyFlag: added support for level bounds
• netCDF: added support for time axis name and long_name
• cgribexGetGrid: bug fix for xinc/yinc recomputation
• stream_cdf::define_all_vars: added txt attributes to vlistDefAttTxt() without trailing 0 [Bug #3004]

New Feature:

• new return option: Masked Arrays
  if the new keyword returnMaArray is given, its value is taken as variable
  name and a masked array wrt to its FillValues is returned, examples for python and ruby can be found in the unit tests
  contribution for python by Alex Loew
• Error handling: return stderr in case of non-zero return value + raise exception
  contribution for python from Estanislao Gonzalez (#3047)
• Autocompletion and built-in documentation through help() for interactive use
  contribution from Estanislao Gonzalez [python] (#3048)
• Added help operator for displaying help interactively [ruby]
  "sample": https://github.com/Try2Code/cdo-bindings/blob/master/ruby/test/test_cdo.rb#L299

Installation works as usual with pip and gem.

New features:

• Ruby and Python bindings now use the same keys to define input and output streams
• New option force to suppress reprocessing existing output files
• New option returnArray for return a numpy/narray for a given variable name

Please have look at the updated documentation at cdo.{rb,py}

Redmine supports git repositories, but to use them from another server is not straightforward. A new wiki entry in Guidelines tells you as project manager how things work.

• Added support for netCDF4(HDF5) formatted SCRIP grid description files
• added CDO option -L to lock all I/O calls. This option is neccessary if external I/O libraries like netCDF4 (HDF5) were installed without thread-safe support.

• setunit: Set variable unit
• chunit: Change variable unit

• Info: changed output format
• Sinfo: changed output format
• Diff: changed output format

• remaplaf: fixed bug in binary_search_int()
• eca_rr1: result has wrong long name attribute

• grib2: added stepType support for absolute time axis
• netCDF: set 1D arrays to coordinate variables if axis attribute is available

• grib2: changed packingType from grid_jpeg to grid_simple if nmiss > 0

• Wrong results with the following statistical functions: * mean, *avg, *sum, *var, *std
  only if all of the following conditions are complied:
  - x86_64 machine (tornado, squall, thunder, lizard)
  - dataset has no missing values
  - the horizontal grid size is > 1 and not multiple of 8
  This bug was introduced in CDO version 1.5.6.

• Runstat: OpenMP parallelization over parameter nts
• import_binary: added support for 64-bit floats via extra OPTION keyword flt64

• Arith: added support for 3D masks
• mastrfu: use grid coordinates from input file

• ymonsub: added support for time bounds
• Wrong netCDF output for unscaled uint8, int8, int16, int32 variables [Bug #2516]

• netCDF: added support for environment variable NC_CHUNKSIZEHINT
• GRIB2: added support for level type HYBRID_HALF
• GRIB2: added support for grib_api key stepType with TSTEP_<type> in vlistDefVar()

• wrong netCDF output for unscaled uint8, int8, int16, int32 variables (bug introduced in 1.5.5)