[netCDF #KRG-285946]: nc_evarsize

> I've just looked at the one explanation I found for an nc_evarsize and
> still don't understand what my problem may be.  I hope you have some ideas.
>
> Scenario:
> I have a created a netCDF dataset with several 2D nf90_real variables.
> ...
> nf_90_redef(...)
> < nf90_def_dim (..) for a third dimension>
> then   <nf90_def_var(...); nf90_put_att(...)>   for 8 real variables
> nf90_endDef(...)
>
> And, as you know, it is on the endDef call that the nc_evarsize error is
> returned.  All teh real data have nf90_real in their nf90_put_att call.
> None of the earlier netCDF calls on this dataset returned a non-zero
> error code.
>
> Thank you for any help or guesses about my problem.
>
> Lee
>
> --

Howdy Lee!

I hope all is well up there in Alaska!

As for the EVARSIZE error, it occurs when you define a variable that is too
large for the netCDF file format you are using. For information about netCDF
formats, see:
http://www.unidata.ucar.edu/software/netcdf/docs/netcdf/Format.htmlhttp://www.unidata.ucar.edu/software/netcdf/docs/netcdf/NetCDF-Classic-Format-Limitations.html

To get larger sizes, try using the 64-bit offset format. Use the
NF90_64BIT_OFFSET flag when calling nf90_create to get a 64-bit offset file.
Then your limits increase:
http://www.unidata.ucar.edu/software/netcdf/docs/netcdf/NetCDF-64-bit-Offset-Format-Limitations.html

If these two formats are both too restrictive for your needs, you can try the
netCDF-4 beta release, available on the home page, which adds the HDF5 format
to the two above, which has far fewer restrictions on the sizes of objects in
the file.

Does this answer your questions?

Thanks!

Ed

The default format, netcdf classic, is the most restrictive in terms of size of
variables. See here:


Ticket Details
===================
Ticket ID: KRG-285946
Department: Support netCDF
Priority: Normal
Status: Closed