Open ianthomas23 opened 2 years ago
In the GIS world, I think it makes sense to assume that the data array should contain at least 2 dimensions of (y, x)
, or (lat, lon)
coords. In fact, some of the xarray-spatial algorithms do not take coordinate values into account (zonal tools, multispectral tools for example), but if one does, then the coords should be named correctly IMO.
Following discussion, we will restrict this to (y, x)
coordinates that are each 1d arrays initially.
xarrays can have no
(x, y)
coordinates, or they can have 1d or 2d coordinates. Need to check that all algorithms support all 3 use cases, and add tests for them all.New GPU algorithms certainly don't support this yet.