Previous: Unidata's Common Data Model and NetCDF Java Library API Overview Next: What is Chunking? Table of contents Frames 2010 Unidata NetCDF Workshop

18 Chunking and Deflating Data with NetCDF-4
NetCDF-4/HDF5 data may be written in chunks for increased performance, and may use on-the-fly compression.

18.0  What is Chunking?
Chunked data are written in chunks of user-defined size.
18.1  Choosing Chunksizes
How do you pick chunksizes?
18.2  Setting the Chunksizes in NetCDF-4
Set the chunksizes for variables in netCDF-4 file with nc_def_var_chunking().
18.3  Example of Setting Chunksizes in Fortran 90
In this example code (from nf_test/f90tst_vars.f90) the chunksize is set for a 2-dimensional variable in Fortran 90.
18.4  The Cache in NetCDF-4
The cache is used when reading or writing data. Set the cache size correctly for each variable to improve performance.
18.5  Setting the Cache in NetCDF-4
Set the cache for a variable with the nc_var_set_cache function.
18.6  Per-Variable Compression in NetCDF-4
Chunked variables in netCDF-4 files may use on-the-fly compression with zlib.
18.7  Per-Variable Compression in NetCDF-4
This example shows how to set the compression level for variables.
18.8  Per-Variable Compression in NetCDF-4, Results
File sizes of compressed and uncompressed 2D radar data.
18.9  How to Use Per-Variable Compression in NetCDF-4
Some advice about using compression.
18.10  Contiguous Variables
A contiguous variable is stored as one long array in the file.
18.11  Using Contiguous Variables
Use the nc_def_var_chunking function to make a variable contiguous.
18.12  Example of Contiguous Variable
A contiguous storage variable is created in this example.

 


Previous: Unidata's Common Data Model and NetCDF Java Library API Overview Next: What is Chunking? Table of contents Frames 2010 Unidata NetCDF Workshop