[gdal-dev] gdal-dev Digest, Vol 198, Issue 14

Ari Jolma ari.jolma at gmail.com
Fri Nov 13 00:55:53 PST 2020


jratike80 kirjoitti 12.11.2020 klo 11.16:
> Hi Ari,
>
> I fear that user cannot control this with the switches. This is the reason
> for the error:


Yes, looks a bit like the logic for getting a small piece of the data 
that GDAL does fails in this case. I can take a look once I have time.

Ari


>
> ```
> WCS: Requesting
> http://demo.geo-solutions.it/geoserver/wcs?SERVICE=WCS&REQUEST=GetCoverage&VERSION=2.0.1&COVERAGEID=nurc__mosaic&SUBSET=Long(6.3953399152693953,6.3953399152693953)&SUBSET=Lat(46.541601968340089,46.541601968340089)&Format=image/tiff
>
> ```
>
> Thus GDAL creates subsets with lower limit = upper limit which makes the
> bbox into a point. That it probably not intentional. I think it is done with
> the code that begins from
> https://github.com/OSGeo/gdal/blob/master/gdal/frmts/wcs/wcsdataset.cpp#L562
>
> If theory interests, there has been discussions within the OGC WCS group
> about the validity of such request. It may be that the server should expand
> the bbox and send tiff with one pixel as an output. If the subsets were
> written for slicing
> SUBSET=Long(6.3953399152693953)
> &SUBSET=Lat(46.541601968340089)
> then tiff as outputformat is not possible because image has always  two
> dimensions but here 0 dimensionality (point) is requested. But this is a
> sidepath because I believe that the WCS driver wants to read a small area
> from WCS as a sample, not just one pixel.
>
> -Jukka-
>
>
> Ari Jolma-2 wrote
>> 2) gdalinfo
>> "WCS:https://demo.geo-solutions.it/geoserver/wcs?version=2.0.1&coverage=nurc__Img_Sample"
>>
>> .. takes again some time as it makes calls and then it stores new data
>> into the cache
>>
>> -
>> cJLrY=https://demo.geo-solutions.it/geoserver/wcs?version=2.0.1&coverage=nurc__Img_Sample
>> - three files: cJLrY.xml, cJLrY.xml.aux.xml, cJLrY.DC.xml, the first is
>> the WCS_GDAL XML
>>
>> and prints
>>
>> ERROR 1: HTTP error code : 404
>> ERROR 1: InvalidSubsetting: Empty intersection after subsetting
>>
>> so it probably needs some special switches, which are described in the
>> documentation.
>>
>> Best,
>>
>> Ari
>
>
>
>
> --
> Sent from: http://osgeo-org.1560.x6.nabble.com/GDAL-Dev-f3742093.html
> _______________________________________________
> gdal-dev mailing list
> gdal-dev at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/gdal-dev


More information about the gdal-dev mailing list