NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.

To learn about what's going on, see About the Archive Site.

Re: Compressing netcdf files


blincoln wrote:
> 
> On Mon, 07 Aug 2000 16:20:36 -0600, John Caron wrote:
> 
> >this is very impressive; i wonder if this file happens to have a lot of
> >zeroes or missing data or something?
> >
> >floating point is known to be difficult to compress due to the fact the
> >low order bits are essentially random.
> 
> I haven't worked with bzip2 on many types of data yet, but I have
> been thoroughly impressed with it in several areas, compressing
> dramatically better than zip, compress, or gzip in about the same time
> as gzip (somewhat slower than zip) and decompressing pretty quickly.
> 
> Although when I tried it on a couple of several MB netCDF files I had
> (arps runs) I didn't see any real improvement over gzip or zip,
> I'd say its definitely worth trying if you have any in-house
> compression needs.. it isn't heavily used yet, but I have seen lots
> of systems start to use it in the last couple months because of the really
> substantial improvement in compression for some applications.
> 
> bcl

The compression ratios must be data dependent, but bzip2 is clearly
better than others on common netcdf files.  Russ sent this result for
some of our ldm files:

 size  compress  gzip  bzip2
 31Mb   2.54     4.13  6.04
  4Mb   3.63     4.47  7.82
 16Mb   3.58     4.32  7.36

  • 2000 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the netcdfgroup archives: