[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: 20010731: netCDF file size

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: Re: 20010731: netCDF file size
  • Date: Thu, 02 Aug 2001 11:21:08 -0600
>From: "Reggie Chang" <address@hidden>
>Subject: file size
>Organization: ZONA Technology
>Keywords: 200107310123.f6V1NH117343 netCDF file size

Hi Reggie,

> I'm trying to convert the data file (unformatted,
> written out from FORTRAN code) into NetCDF 
> format.  Although I have the results correct, I 
> notice that the file size in the NetCDF format is 
> a lot bigger than original one.

In general, the overhead for netCDF files is low; for details see the
description of the format in Appendix B of the User's Guide:

 
http://www.unidata.ucar.edu/packages/netcdf/f90/Documentation/f90-html-docs/guide67.html#1017343

Netcdf files need not be any larger than other binary formats such as
unformatted binary files written out from Fortran, if the external
netCDF types match the Fortran types you are using.  But if you're
writing doubles for Fortran REAL variables on platforms where REALs
are stored in 32-bits, for example, you should expect the file to be
twice as large, since you're using 64-bit doubles to store 32-bit
values.

To say anything much more specific, we would have to see the CDL for
the file (the output of "ncdump -h foo.nc").

> I found a tip in NetCDF User's Guide for FORTRAN.
> Basically it is related to NF_ENDDEF/NF_REDEF.
> (Chapter 9)  Is this an overhead to execution or file 
> size?  Will arrays with 'unlimited' time coordinate
> cost anything?

That's about execution time, not file size.

There is only one unlimited dimension in a netCDF file, and all the
arrays that use it have the same size in that dimension.  So if most
of your arrays have 10 times but you've written the 100th time for one
array, then all of the arrays will be extended to have 100 times.

But arrays that use the unlimited dimension don't cost any extra size
just because they use the unlimited dimension.  In fact if you haven't
written any data for such variables, the current size of the unlimited
dimension will be zero and the arrays won't take up any space in the
file.

--Russ

_____________________________________________________________________

Russ Rew                                         UCAR Unidata Program
address@hidden                     http://www.unidata.ucar.edu