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

[netCDF #KLN-901251]: netcdf-4 with C++ interface in the future?

Hi Kevin,

> I attended the netCDF workshop this year and really got a lot out of
> it. One of the main things I wanted to bring back to NSSL was how we
> might be able to incorporate the netCDF-4 read/write capabilities in our
> WDSSII API which is C++ based.
> 
> I, unfortunately, am unclear about an important issue when reviewing my
> notes from the workshop. As I recall, the C++ interface lags behind
> some of the other implementation, particularly in the ability to write
> netCDF-4 files. I see the webpage
> 
> <http://www.unidata.ucar.edu/software/netcdf/docs/cxx4/> that
> 
> regarding the Appel rewrite of the interface and - as I understand it -
> this is included in the Unidata distribution. When looking through the
> mail archives, I found this message "[netCDF #KLN-901251]: netcdf-4 with
> C++ interface in the future?" and became confused as to whether this
> implementation can write netCDF-4 files that use the classic data model
> format:
> 
> "The limitation in the current implementation that needs to be
> addressed before we announce the availability of this contributed
> software is its inability to create netCDF classic or netCDF-4
> classic-model format files. Currently it can only create netCDF-4
> files."
> 
> Can you confirm for me whether or not the C++ interface for netCDF-4 (v.
> 4.1.1 or 4.1.2-beta) can read and write netCDF-4 files in the classic
> data model?

The version we are currently distributing cannot write netCDF-4 files
in the classic data model, in the sense that they are not "marked" by
a special internal attribute as being netCDF-4 classic files, but it
can create netCDF-4 files that are otherwise classic model files.

It can read all 4 kinds of netCDF files, including classic, 64-bit
offset, netCDF-4, and netCDF-4 classic-model.

Modifying the NcFile constructor to properly interpret the kind of
file to create and pass it on to the underlying C interface should be
relatively simple, but has not been done yet.  If we don't have the
modifications needed to do this after I return from the AGU meeting in
a couple of weeks, I may dust off my rusty C++ skills and try to get
it done.  If you need it sooner than that, your C++ developers might
be able to polish it off and contribute the changes back to us.

--Russ

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



Ticket Details
===================
Ticket ID: KLN-901251
Department: Support netCDF
Priority: Normal
Status: Closed