[THREDDS #ETD-820941]: aggregation and datasetscan

• To: venus@xxxxxx
• Subject: [THREDDS #ETD-820941]: aggregation and datasetscan
• From: "Unidata THREDDS Support" <support-thredds@xxxxxxxxxxxxxxxx>
• Date: Wed, 13 Dec 2006 13:19:28 -0700

Hi Valentijn,

> The example seems not to work (see Shawns report as attached).
>
> > I tried and this is what I got! (figures below, fyi)
> >
> > It worked fine with catalog. The problem occurs when the client
> > starts to $B!H(Jread$B!I(J the dataset. It may due to 1). the
> > incorrect aggregation of grid or 2). The grid parsing capability
> > on the client, my guess.

What client are you using to access the data? IDV? ToolsUI? Something else? (If 
this last one, does it use the netCDF-java library?)

I'd like you to try two things.

1) Pull the entire netcdf element out of the catalog and store it in a local 
file with a ".ncml" extension. Actually, I'll just attach a file to this 
message. Open the file with the IDV or ToolsUI. If in IDV, choose "Grid files" 
from the "Data Source Type" pull down. If in the ToolsUI, open in the "Grids" 
tab.

2) Back to the TDS served catalog. First, make sure the catalog matches the 
dataset ncml that works above. Then send me the response you get to these two 
URLs:

- http://localhost:8080/thredds/dodsC/valentijnAggOfUNH.dds
- http://localhost:8080/thredds/dodsC/valentijnAggOfUNH.das

> I was hoping we could use the crawlableData for this so as
> to bypass the need for the "location" tag. This would save
> a lot of time if it concerns a very lengthy list of datasets.

Yes, that is the plan for down the road. However, the aggregation scan is the 
part we've talked about before that does not use the CrawlableDataset code. It 
uses some seperate file-only code for scanning. It is on our list to change 
this to use the CrawlableDataset code but it is not there yet.

Ethan

Ticket Details
==================
Ticket ID: ETD-820941
Department: Support THREDDS
Priority: Critical
Status: Open