Per the discussions, can you please also add a test to check that pending trials that don't yet have results, still show up in the results_df, with empty results columns?
Originally posted by @bpkroth in #889 (comment)
Given the way that the _dummy_run_exp works in tests/storage/sql/fixtures.py this is probably easiest to do if we first add the API in the storage layer for enqueuing a new trial manually.
See Also: #687