What would you like to be improved?

In setUp() of TestHadoopCatalogOperations, we stub schema files in a fixed‐length loop:

As we add new tests that require additional schemas to be stubbed (e.g. schema30, schema31, ...) , this manual update of the loop's upper bound is inefficient and throws misleading test failures.

How should we improve?

Move the schema-stubbing action to its corresponding test.

cc @mchades