Packed Data Values

11.8 Packed Data Values

Conventions for packed netCDF data use a smaller data type to save space for low-precision data, for example, packing doubles into shorts.

The C-based netCDF library itself does not do packing and unpacking, but the Java-netCDF library will do automatic unpacking, see documenation for the class VariableDS.

Recommendations

For details, see the Packed Data Values section of the netCDF Best Practices document. There are some subtleties involving avoiding introducing bias when packing, and how to represent missing values.

2011 Unidata NetCDF Workshop > Best Practices