Issues with dimension names in netcdf files

[danielbwright]

Currently, the user can specify the variable names for the x and y coordinates, e.g., "latitude", "longitude"). However, if the dimension names don't match the variable names, readnetcdf() will throw an error here: outrain=infile[rain_name].sel(**{lat_name:slice(latmin,latmax)},\ **{lon_name:slice(longmin,longmax)}) But really 'lat_name' and 'lon_name' are the dimension names, not the variable names. Can we make the code robust to this inconsistency in variable and dimension names?

[danielbwright]

As an alternative, we can provide very explicit instructions that the variable names and the dimension names must be identical.

[Ashar25]

Yes, the dimension name should be identical with the coordinates name. The purpose of asking variable is to handle the variability we get from different datasets like some datasets name their X and Y as: "Lat", "Lon" or just X and Y and the precipitation variable as: prep or rain or whatever. Users can mention the name of these "variables" (either coordinate or the precipitation) in the JSON file and the program can handle that.