You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
If you use a Grid_R with lon and lat dimensions of the same length, and data ordered as (lat, lon), the streams will get crossed when you interpolate and your output gets mirrored.
In the case where they are the same size, further attribute cross referencing needs to be done to ensure correct dimension association
I think this should be embedded n the CF standard.
As a rule, the code should always try to use the "standard" way to assign stuff, and only infer things is it has to.
This may be a case where it has to be inferred, but we should make that's so. ANd if it is, well, the fie is non compliant, and you get what you get :-)
Looks like we already had this bug reported, but it bit us again with the files Chris created recently. We need to prioritize a fix on this or disable loading regular grids with the same dimensions in lat/lon in PyGNOME before it gets us yet again...
NOTE: we should be able to use the shared dimensions to figure out which is lat and lon.
i.e. the longitude variable should use a dimension that is different than the latitude variable, and by looking at the dimensions used by the other variables, know which to use.
If you use a Grid_R with lon and lat dimensions of the same length, and data ordered as (lat, lon), the streams will get crossed when you interpolate and your output gets mirrored.
In the case where they are the same size, further attribute cross referencing needs to be done to ensure correct dimension association
gridded/gridded/grids.py
Lines 356 to 357 in 5217b8f
The text was updated successfully, but these errors were encountered: