Re: [Fwd: Re: THREDDS and grib]

[Ethan Davis <edavis@xxxxxxxxxxxxxxxx>]

So the IOSP translate this as two separate Groups, one for each projection.

John, is there a way to aggregate on the dimension in one of the groups? 
I just tried

<?xml version="1.0" encoding="UTF-8"?>
<netcdf xmlns="http://www.unidata.ucar.edu/namespaces/netcdf/ncml-2.2";>
 <aggregation dimName="proj0.time" type="joinExisting">
   <netcdf location="file:Z:/kevinGribData/godas.P.19800105.grb" />
   <netcdf location="file:Z:/kevinGribData/godas.P.19800110.grb" />
   <netcdf location="file:Z:/kevinGribData/godas.P.19800115.grb" />
 </aggregation>
</netcdf>

but that didn't work. I got

java.lang.IllegalArgumentException: Dimension length =0 must be > 0
   at ucar.nc2.Dimension.setLength(Dimension.java:246)
   at ucar.nc2.Dimension.<init>(Dimension.java:192)
   at ucar.nc2.Dimension.<init>(Dimension.java:166)
   at 
ucar.nc2.ncml4.AggregationExisting.buildDataset(AggregationExisting.java:70)
   at ucar.nc2.ncml4.Aggregation.finish(Aggregation.java:249)
   at ucar.nc2.ncml4.NcMLReader.readNetcdf(NcMLReader.java:407)
   at ucar.nc2.ncml4.NcMLReader.readNcML(NcMLReader.java:360)
   at ucar.nc2.ncml4.NcMLReader.readNcML(NcMLReader.java:203)
   at ucar.nc2.ncml4.NcMLReader.readNcML(NcMLReader.java:153)
   at ucar.nc2.dataset.NetcdfDataset$1.open(NetcdfDataset.java:315)
   at ucar.nc2.NetcdfFileCache.acquire(NetcdfFileCache.java:205)
   at ucar.nc2.dataset.NetcdfDataset.acquireFile(NetcdfDataset.java:313)
   at ucar.nc2.ui.ToolsUI.openFile(ToolsUI.java:799)
   ...

Also, the OPeNDAP server shows duplicate variables in the DDS. What is 
the plan for serving datasets with Groups and such? Or is this fixed in 4.0?

Ethan

Robb Kambic wrote:
> On Wed, 5 Mar 2008, Ethan Davis wrote:
>
> > Hi guys,
> >
> > Kevin at PMEL is trying to aggregate some GRIB files and running into 
> > problems.
> >
> > I have some sample files at /home/edavis/kevinGribData.
> >
> > When I open them up in ToolsUI, I see a proj0 and proj1 structure. 
> > They each have a latLonCoordSys variable but I don't see a difference 
> > between them.
>
> Ethan,
>
> there is a minor differance in the projections, ie p1.Lo1 != p2.Lo1. 
> That's enough to cause problems in an aggregation.
>
> --------------------------------------------------------------------
> GDSkey = 2494715761
> La1 = -74.5
>
> Lo1 = 0.5
>
> La2 = 64.499
> Lo2 = 359.5
> --------------------------------------------------------------------
> GDSkey = 2548847210
> La1 = -74.0
>
> Lo1 = 1.0
>
> La2 = 64.999
> Lo2 = 360.0
>
>
> Before the aggreagtion could occur, some of the grid parameter would 
> have to be reprojected to the other grid co-ordinates or the 
> parameters would have to be seperated out into files according to 
> co-ordinates and then aggregated.
>
>
> Robb...
>
>
>
>
> I'm
> > guessing this is confusing the aggregation. Are there actually two 
> > different projections in each file or is the GRIB library or IOSP 
> > mapping things poorly?
> >
> > Could you take a look at these files and let me know what is up?
> >
> > Thanks,
> >
> > Ethan
> >
> > --
> > Ethan R. Davis                                Telephone: (303) 497-8155
> > Software Engineer                             Fax:       (303) 497-8690
> > UCAR Unidata Program Center                   E-mail:    edavis@xxxxxxxx
> > P.O. Box 3000
> > Boulder, CO  80307-3000                       
> > http://www.unidata.ucar.edu/
> > --------------------------------------------------------------------------- 
>
> =============================================================================== 
>
> Robb Kambic                       Unidata Program Center
> Software Engineer III               Univ. Corp for Atmospheric Research
> rkambic@xxxxxxxxxxxxxxxx           WWW: http://www.unidata.ucar.edu/
> =============================================================================== 

--
Ethan R. Davis                                Telephone: (303) 497-8155
Software Engineer                             Fax:       (303) 497-8690
UCAR Unidata Program Center                   E-mail:    edavis@xxxxxxxx
P.O. Box 3000
Boulder, CO  80307-3000                       http://www.unidata.ucar.edu/
---------------------------------------------------------------------------