python-xarray：单维数据的多维坐标

Question

I have a number of datasets in xarray that I eventually would like to combine into one. One such dataset could have, for example, a datarray named 'x' , which has dimensions 'a' , 'b' and 'c_d_e' . I have corresponding parameters that were used to generate the data, and I would like to use these as coordinates. In case of the dimension c_d_e , it is actually three different parameters that were varied together, ie only one dimension of data is necessary for them.

So while my dimensions are ['a', 'b', 'c_d_e'] , I would like my coordinates to be 'a', 'b', 'c', 'd', 'e' , and later combine datasets along these dimensions.

Code example (of what I would like):

xr.Dataset({'x': (['d1', 'd2', 'd3_d4_d5'], x), coords=['d1coord':d1arr, 'd2coord': d2arr, 'd3coord': d3arr, 'd4coord': d4arr, 'd5coord': d5arr])

I could of course use all the 'a', 'b', 'c' as dimensions, but that would make xarray fill the array with NaNs in places that don't correspond to and combination of coordinates, greatly bloating the amount of space that this takes up.

Answer 1

Okay, I have found a solution, and it is so stupidly simple that I cannot believe that I didn't think of that earlier.

I simply specify one of the variables 'c', 'd', 'e' as a dimensional variable and the others as non-dimensional (they are just a function of the former).