With a local athena equivalent db available, we should be able to use it instead of mocked pyathena cursors, which would be a better test of actual behavior (although we will catch that with regression tests).
We should make a fixture to handle spinup/teardown, and then use that fixture in each test that currently uses a mocked DB curosr.
With a local athena equivalent db available, we should be able to use it instead of mocked pyathena cursors, which would be a better test of actual behavior (although we will catch that with regression tests).
We should make a fixture to handle spinup/teardown, and then use that fixture in each test that currently uses a mocked DB curosr.