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

RE: 19991202: Java NetcdfFile.add

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: 19991202: Java NetcdfFile.add
  • Date: Fri, 3 Dec 1999 17:41:32 -0700
>
> > Hmmm, unfortunately, the Java netcdf API doesnt allow you to
> > incrementally update the Schema; the Schema in a NetcdfFile is
> > considered immutable.
>
> Any chance the java NetCDF could be extended in this direction?
>

Possibly eventually, but not immediately

> Also support for ragged arrays would be greatly appreciated.

we are considering using the hdf-5 data file format underneath our next
version of netcdf (version 4). You might look at that now; theres no
Java version though, and we arent yet decided on that path.

> >
> > Kinda lame, but I think thats the way it is. You could 1)
> read the data
> > twice, accumulating the scheme on the first pass, or 2)
> write seperate
> > files for each Variable, then merge at the end.
>
> How do you merge netcdf files? It seems like the same problem.

You can read in the metadata from the single-variable files without
reading the data, construct the schema, make the big file, then read the
data and fill the big file.

>
>
> I have my code working. I simply store the data locally as I build
> the schema and then read back through the data to write it to the
> file. This will work fine for modest data set sizes but for larger
> datasets this will be a problem.
>

sounds like you did the equivalent of what i just suggested. extra cost
is one local read and the extra scratch disk space.