Reprojection results in nan array

I have been trying to reproject DECaLS (fk5) data to SDSS (ICRS) data but either way it results in a matrix with all nan values with no error. I dont know what I am doing wrong or if its a limitation with reproject.

No source file, test, or entry point is named. Start by reproducing the all-NaN result with the reported DECaLS and SDSS inputs, then inspect the reprojection behavior and coordinate assumptions. Done means determining whether the result is caused by the inputs or a library limitation, with a documented resolution or a focused fix and regression test.

Written by the indexing model from the issue text.