[gdal-dev] array dim permute

[Even Rouault]

Hi Michael,

I guess this is finally the perfect use case we were waiting for to find 
a good reason to embed a Small Language Model within GDAL to figure out 
what are the columns and lines axis.  In the meantime, the boring 
https://github.com/OSGeo/gdal/pull/12700 will fix it

Even

Le 04/07/2025 à 00:48, Michael Sumner via gdal-dev a écrit :
> I'm confused about the orientation of this in 2D mode 1485x1272 is 
> correct, with 122 bands.
>
> (requires earthdata auth)
>
> gdal raster info 
> "/vsicurl/https://obdaac-tea.earthdatacloud.nasa.gov/ob-cumulus-prod-public/PACE_OCI.20250501T234515.L2.SFREFL.V3_0.nc" 
>   --output-format text | grep /rhos
> # 
>  SUBDATASET_1_NAME=NETCDF:"/vsicurl/https://obdaac-tea.earthdatacloud.nasa.gov/ob-cumulus-prod-public/PACE_OCI.20250501T234515.L2.SFREFL.V3_0.nc":/geophysical_data/rhos
> #  SUBDATASET_1_DESC=[1485x1272x122] /geophysical_data/rhos (32-bit 
> floating-point)
>
> If we drill into the rhos sds X becomes unrolled as the 1485 bands:
>
> gdal raster info 
> "vrt:///vsicurl/https://obdaac-tea.earthdatacloud.nasa.gov/ob-cumulus-prod-public/PACE_OCI.20250501T234515.L2.SFREFL.V3_0.nc?sd_name=/geophysical_data/rhos" 
> --output-format text | grep "Size"
> #Size is 122, 1272
>
> We can (redundantly??) cast via mdim to get 1485x1272, but then we've 
> lost the geolocation arrays:
>
> gdal raster info 
> "vrt:///vsicurl/https://obdaac-tea.earthdatacloud.nasa.gov/ob-cumulus-prod-public/PACE_OCI.20250501T234515.L2.SFREFL.V3_0.nc?transpose=/geophysical_data/rhos:0,1" 
> --output-format text | grep "Size"
> #Size is 1485, 1272
>
> At any rate in the default case the geolocation arrays don't match the 
> data orientation.
>
> Cheers, Mike
>
> -- 
> Michael Sumner
> Ordinary Member,  Streets People Love Hobart Association
> Research Software Engineer
> Australian Antarctic Division
> Hobart, Australia
> 0438489030
> e-mail: mdsumner at gmail.com
>
> _______________________________________________
> gdal-dev mailing list
> gdal-dev at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/gdal-dev

-- 
http://www.spatialys.com
My software is free, but my time generally not.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/gdal-dev/attachments/20250704/77921de2/attachment.htm>