Hi Andrew,
You can configure the THREDDS Data Server (TDS) to generate the currently
available datasets 'on the fly'. The main docs for the TDS are at
http://motherlode.ucar.edu:8080/thredds/docs/
The "Getting Started" link
(http://motherlode.ucar.edu:8080/thredds/docs/GettingStarted.html) goes over
the basics. What you probably want is to include a <datasetScan> element in the
TDS config file. There is details on the datasetScan at
http://motherlode.ucar.edu:8080/thredds/docs/datasetScan/
_From your example, something like this might work:
<?xml version="1.0" encoding="UTF-8"?>
<catalog
xmlns="http://www.unidata.ucar.edu/namespaces/thredds/InvCatalog/v1.0";
name="Unidata IDV Catalogs">
<service name="champ" serviceType="OPeNDAP" base="/thredds/dodsC/" />
<dataset name="my data">
<metadata inherited="true">
<serviceName>champ</serviceName>
</metadata>
<datasetScan name="CHAMP netCDF" path="champ"
location="/usr/local/AMPS/idv/data/incoming/champ/nz12kmn/" />
</dataset>
</catalog>
You might want a seperate datasetScan for champ and one for ec_ens. I wasn't
sure what the directory structure looked like so I only did one datasetScan.
Also, I only included an OPeNDAP server. You can include other data services as
described in the TDS Configuration reference:
http://motherlode.ucar.edu:8080/thredds/docs/Configuration.html
I'm not sure about including the NcML file. I will leave the answer to this for
John to answer.
Hope this helps.
Ethan
> I am having a bit of trouble setting up a THREDDS server, however. I'm
> able to create a THREDDS catalog which points to explicitly named files
> - see below. What I'd like instead is to provide links to the newest
> file in a directory, and the next-newest, etc, or just provide a list of
> all files. Is this possible?
>
> I have a directory structure that looks something like this:
>
> champ/2006100806.nc
> champ/2006100812.nc
> champ/2006100818.nc
>
> ec_ens/2006100806.nc
> ec_ens/2006100812.nc
> ec_ens/2006100818.nc
>
> ..etc
>
> Can I setup the THREDDS catalog to generate a list of the files in these
> directories 'on the fly', or do the files have to be specified
> explicitly in the dataset specification of the THREDDS file - ie. would
> I have to re-create the xml file every time the files in the directory
> changes? I've tried to work this out using the examples in the Catalog
> data chooser, but these all seem to have the source files specified
> explicitly.
>
> Also, how do I specify the NcML file in the THREDDS catalog?
>
> Sorry to keep bothering you with these sorts of things, but we're almost
> there!
>
> Thanks for your help,
>
> Andrew
>
>
>
>
>
>
> -----------------------------------------------------------
> Current THREDDS xml file:
>
> <?xml version="1.0" encoding="UTF-8"?>
> <catalog
> xmlns="http://www.unidata.ucar.edu/namespaces/thredds/InvCatalog/v1.0";
> xmlns:xlink="/usr/local/AMPS/idv/" name="Unidata IDV Catalogs">
>
> <service name="champ" serviceType="http"
> base="/usr/local/AMPS/idv/data/incoming/champ/nz12kmn/" />
> <dataset name="CHAMP netCDF" >
> <dataset urlPath="2006100818.nc" name="2006100818" serviceName="champ"/>
> </dataset>
Ticket Details
==================
Ticket ID: GFM-139212
Department: Support THREDDS
Priority: Normal
Status: Open
