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

[netCDF #LFC-698377]: Impact of change to netCDF products

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: [netCDF #LFC-698377]: Impact of change to netCDF products
  • Date: Mon, 02 Jul 2012 14:07:11 -0600
Hi Becky,

> I need your input on the impact of a proposed change.  I warn you I know
> very little about netCDF.
> 
> Our modelers put out data from the Global HYCOM data in netCDF.  It's
> available at http://nomads.ncep.noaa.gov/pub/data/nccf/com/rtofs/prod/ .
> They currently put out data on 10 levels (I think it's 10) and they want
> to add more levels to make it 33 levels.
> 
> The question is -- what kind of impact, if any, could this have on our
> user community?  I know there is the increase in filesize.  But in terms
> of decoding and using the data, how much could this blow any of them out
> of the water?  With GRIB we have the issues of users needing necessary
> tables, and the fact that records are in order, so if someone hardwired
> where u wind is, changing that could cause user's trouble.  But I'm told
> that netCDF does not have this type of order, so putting more records in
> it won't cause these problems.

That's right, it shouldn't cause any problems.  That's assuming the extra
levels being added are just more of the same type of levels that are
already there.

> I need to advise my management on what effect this change might have on
> customers, so I'm hoping you can give me your opinion.

It should have no effect if the programs the customers are using determine
the index for a level by its corresponding coordinate variable, instead of
using hard-wired knowledge such as "the 500 HPa levels are always in the 3rd
level".  However, it is possible to write programs that ignore the metadata
supplied, and instead rely on short cuts for what works, without considering
what might change.  The netCDF APIs don't enforce doing the right thing, they
just encourage it by making it easy ...

--Russ

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



Ticket Details
===================
Ticket ID: LFC-698377
Department: Support netCDF
Priority: Normal
Status: Closed