Re: 19990312: mangled doubles, 3.4, NC_UNLIMITED first dim?

Phil,

>Date: Mon, 15 Mar 1999 17:26:34 -0700
>From: Phil S <address@hidden>
>Organization: SNL
>To: Steve Emmerson <address@hidden>
>Subject: Re: 19990312: mangled doubles, 3.4, NC_UNLIMITED first dim?
>Keywords: 199903122213.PAA00700

In the above message, you wrote:

> Thanks for the reply. Just now I got my first glimmer of hope of a fix!
> 
> It turns out that I have been doing what ammounts to this 3-stage
> procedure:
> 
>       create
>               redef
>               dimdef...etc.
>               enddef
>       close
> 
>       open
>               putvar...(with some various NC_UNLIMITED slabs)
>       close
> 
>       open
>               redef
>               dimdef, etc.!
>               enddef
>               putvar
>       close
>       
> Well, AFAICT, netCDF "doesn't like" going back into redef mode after
> it's been writing out record variables and otherwise being finished with
> header-like information. (There are additional intricacies - such as
> multiple units being open multiple times - I'll just gloss over that for
> now. I mention it only in case it might have bearing on future
> replication of the problem.)
> 
> This is barely more than superstition at this point - it seems to be
> fixing my corrupted doubles problem in early preliminary tests. I don't
> know if the problem I was seeing would be observable by others, but,
> FWIW, that's my data point for you. 

Thanks for the information.  I might help.

--------
Steve Emmerson   <http://www.unidata.ucar.edu>