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

Re: Netcdf4 Compound types and Examples

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: Netcdf4 Compound types and Examples
  • Date: Thu, 27 Dec 2007 11:15:21 -0700
> Thanks Russ.  Where do I go to get the recent snapshots?  I am
> working from he beta1 and I don't see the files you mentioned in there.

There are links from the netCDF home page, where it says

  Developers may wish to download daily netCDF-3 snapshot release and
  netCDF-4 snapshot release, ...

and here they are explicitly:

  ftp://ftp.unidata.ucar.edu/pub/netcdf/snapshot/netcdf-daily.tar.gz
  ftp://ftp.unidata.ucar.edu/pub/netcdf/snapshot/netcdf-4-daily.tar.gz

These both contain the same version of ncdump and the same examples, but
the first is the current netCDF-3 daily snapshot release.

> it has been interesting trying to write a wrapper.  I have learned
> more about pointers and structures and C and Fortran than I ever
> wanted to know.  There is a reason I  put the effort into getting
> netcdf4 to work with Absoft.    To do something like the structure
> for the vlen and to be able to pass it back and forth from C , I need
> to have the following:
>
> 1.  Cray style pointers in F90/95
>
> 2. The ability to allocate space to the cray pointer, i.e a malloc
> type command in Fortran
>
> 3.  Support for the sequence attribute in a structure, to guarantee
> memory is in one block.
>
> Fortunately Absoft supports all of these.  If I ever get this to work
> I will let you know - but it will not be portable.

It's our intention that use of vlens from Fortran 90 should be
portable.  That may not be the case for compound types because of
portability problems in Fortran ...

--Russ