Hi Roy,
I believe you could serve each of your stacked GRIB files as an
individual dataset (using a dataset or datasetScan element).
However, using a featureCollection with featureType="GRIB" has the
advantage that it automatically extracts metadata from the file(s) and
places it in the catalog. I believe the GRIB featureCollection should
work both for aggregating GRIB files or for serving a single stacked
GRIB file. Other than needing a featureCollection element for each
stacked GRIB file, I don't think there are any downsides to going this
route.
I hope that answers your question.
Ethan
On 2/7/2013 11:18 AM, Roy Mendelssohn - NOAA Federal wrote:
> Hi All:
>
> Well, due to some changes at FNMOC, I may have to start serving some
> Grib files directly in TDS. In looking over the docs for TDS 4.3
> given that I want time aggregation, it is clear what option I need to
> use if I have separate grib files. However, Grib files have a
> feature that we use a lot, that you can stack them into one file.
> Grib readers read them just fine. So the question is if I have a
> grib stack do I need to define an aggregation, or are the times read
> automatically?
>
> Thanks,
> -Roy
--
Ethan Davis UCAR Unidata Program
edavis@xxxxxxxxxxxxxxxx http://www.unidata.ucar.edu
