[Support #LRP-531622]: netCDF4 and variable length arrays in height or range

Hi Barb,

> We looked at netCDF years ago and found the only extensible
> parameter was time.  We have sparse data sets in the CEDAR
> DB with really variable altitude or range arrays, and did not
> see how to get around finding the longest array to define range
> or height.  We currently define the range or height at each time
> step.  Is this variable length array problem solved in netCDF4?
> (or netCDF3?)

Yes, netCDF-4 supports definition and use of variable-length types
for data.  Their use is documented starting here:

  http://www.unidata.ucar.edu/netcdf/docs/netcdf-c.html#Variable-Length-Array

> Are netCDF4 and HDF5 converging?  We think HDF5 does not have a
> problem with variable length arrays.  We are trying to decide
> if we should convert our entire 32-bit integer DB to netCDF4
> or HDF5 or whatever.

NetCDF-4 provides many of the useful features of HDF5, because it uses HDF5
as a storage layer.  NetCDF-4 is intentionally simpler than HDF5, so it will 
never converge to HDF5, but the simplicity is intended to be a feature.  A 
list of the differences between netCDF-4 and HDF5 is here:

  http://www.unidata.ucar.edu/netcdf/docs/faq.html#fv15

--Russ

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



Ticket Details
===================
Ticket ID: LRP-531622
Department: Support netCDF
Priority: Normal
Status: Closed