Closed mattdutson closed 5 years ago
After some thought I think the best way to do this is through the Python API. There's likely a module which reads an SQLite table into a Numpy array, which we can then compare to Hustle output after calling Relation.to_numpy()
.
Currently our tests are limited to those which check a single value in an aggregated column. We need a way to compare Hustle results to SQLite results element-for-element while also accounting for the fact that the order may not be the same.