• The netCDF classic file format is mature and implementation is efficient for many uses.
• With netCDF, all file metadata is read into memory when file is opened, so functions that get metadata ("_inq_" functions) are very fast.
• To avoid copying data when the file schema changes, do one of the following
• Create all needed dimensions, variables, and attributes before writing data, or
• Reserve extra space in the file header for later additions (using nc__enddef() in C, NF__ENDDEF() or NF90__ENDDEF() for Fortran, setExtraHeaderBytes() method of NetcdfFileWritable for Java)
• nc__enddef() in C
• NF__ENDDEF() or NF90__ENDDEF() for Fortran
• setExtraHeaderBytes() method of NetcdfFileWritable for Java
• Use netCDF-4 format

2011 Unidata NetCDF Workshop > Formats and Performance