Re: [thredds] NetcdfSubset Not Selecting Entire Spatial Subset with TDS v4.3.16

Hi Michael,

The difference in this case is that in 4.2 you are using an option available in the interface that allows to request the whole horizontal grid and ignores the bounding box in the request. If you check the "bounding box" radio button, then the service will try to subset to the specified bounding box. 4.3 works slightly different and does not have that option available on the interface but if you want to request the whole horizontal grid you just don't pass the parameters that specify the bounding box in the request so with a simple request like this:
<http://localhost:8080/thredds/ncss/grid/curvilinear/archv.2011_003_00_2d.nc?horizStride=1&time_start=2011-01-03T00%3A00%3A00Z&time_end=2011-01-03T00%3A00%3A00Z&timeStride=1&accept=netcdf&var=ssh>http://ncss1.hycom.org/thredds/ncss/grid/GLBa0.08/expt_90.9?time_start=2011-01-03T00%3A00%3A00Z&time_end=2013-05-12T00%3A00%3A00Z&var=ssh
it should return the whole grid and the dataset is only subsetted by vars and time. Actually, you only need to specify the parameter var in 4.3.
Here is the summary of how NCSS params work in 4.3:
https://www.unidata.ucar.edu/projects/THREDDS/tech/interfaceSpec/NetcdfSubsetService_4_3.html#Common

Also, be aware that global datasets and requests crossing the longitude seam are not well supported by NCSS and when you try a request that crosses the seam you'll only get the data on one side of the seam. Besides, the longitude axis in your dataset is denormalized and its values range from 74.1199 to 1019.1199 so when NCSS tries to guess what are the min and max longitude it normalizes the values and gets 74.1199 and 299.1199. Then, it computes the indices in the data arrays for the bounding box with that longitude range and this is why you don't get all the data in your 4.3 request.

Cheers!

On 05/09/2013 02:53 PM, Michael McDonald wrote:
Using the latest TDS v4.3.16, our NetcdfSubset service is not
selecting the entire Global domain (lat/lon subset, bounding box). The
'east' bounding box is getting a different value from our NetCDF files
compared with the previous TDS v4.2.10 NCSS. See figure attachments.

How it used to look:
"TDSv4.2.10 ncss form.png"
"TDSv4.2.10 full domain.png"

How it looks with the latest:
"TDSv4.3.16 ncss form.png"
"TDSv4.3.16 partial domain.png"

I don't want to have to re-write all of our existing NetCDF files. We
might be able to get away with correcting the catalog so that it uses
the proper lat/lon domain.


--
Michael McDonald
Assistant in Research / Systems Administrator
Center for Ocean Atmospheric Prediction Studies
The Florida State University
http://coaps.fsu.edu


_______________________________________________
thredds mailing list
thredds@xxxxxxxxxxxxxxxx
For list information or to unsubscribe,  visit: 
http://www.unidata.ucar.edu/mailing_lists/

  • 2013 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the thredds archives: