NetCDF
- started in 1989
- most used in geoscience community
- array-oriented
- self-describing
header
, describe:- the layout of rest of the file
- arbitrary file metadata in the form of name/value attributes
- machine-independent
- endianness being addressed in the libraries
Formats
- classic format(default)
- variables
- types (char, byte, short, int, float, double)
- dimensions
- attributes
- variables
- 64-bit offset format
- support large variables and sizes
- netCDF4/HDF5 format
- HDF5 data format, with some restrictions
File parts
- The header, containing information about dimensions, attributes, variables, and all attribute values.
- The fixed-size data, containing data values for variables that don't have an unlimited dimension (if there are any)
- The record data, containing data values for variables that have an unlimited dimension (if there are any)
This format has some characteristics important for performance:
- no extra space in the header (by default) or fixed-size data part
- little storage overhead
- a file can grow efficiently only by appending records to the end, growing all record variables together