Re: 970212: ncvarinq

>To: address@hidden
>From: Gregory Roland Ruetsch <address@hidden>
>Subject: ncvid problem
>Organization: Stanford
>Keywords: 199702130101.SAA08277

Hi Greg,

> A have a question regarding how to determine if a variable is present
> in a netcdf file.  I want to postprocess a group of files, some of 
> which contain a variable 'theta' and some of which don't.  I would
> like to determine this at run time to avoid setting flags in the code
> and recompiling.  I have the following code segment:
> 
>   integer ncid, vid, rcode, scalar_field
>  ... 
> 
>   ncid = NCOPN('nc.ic', NCNOWRIT, rcode)
>   vid = NCVID(ncid, 'theta', scalar_field)
>   call NCCLOS(ncid, rcode)
> 
>   if (scalar_field == 0) then
> 
> 
> so that I am using the error parameter in the NCVID call to determine
> if 'theta' exists in the file.  However, when I run the code I get the
> following error:
> 
>    ncvarid: variable "theta" not found
> 
> and the code quits.  Is there a better (one that works) way to do this?

Yes, although the default error handling behavior of netCDF functions is
to exit on error, this behavior is under programmer control. You can
independently control the fatality of errors and the appearance of
messages from errors detected in netCDF library calls. See the section
on Error Handling in the NetCDF User's Guide for how to do this (page 32
in the 2.4 User's Guide) or on-line at:

    http://www.unidata.ucar.edu/packages/netcdf/guide_6.html#SEC37

This explains that you will need to add an initial call to

      CALL NCPOPT(0)

to set the error handling so you can handle all errors.

--Russ

_____________________________________________________________________

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