Improve test coverage, increase code reuse (remove Profile class).
Fix serialization/deserialization issues from yaml 1.1 interpretation of Interval's "on" attribute as a boolean value (even when used as a key).
1.7.0
Remove Persistor's `query_by_keys()` and `df_from_query_by_keys()`. Use `query()` and `df_from_query()` with keyword argument `keys` and `parameters`.
Add caching to Persistor's `df_from_query()` with keyword argument `cache` as the filesystem location to store compressed picked dictionaries of sql -> DataFrame. Caching compares the literal rendered sql only: it is not aware of volatile functions called on the server (like current time) or side effects (like temp tables).
Add rendering of multi-column key values to Persistor's `query()` and `df_from_query()`. Users pass DataFrames instead of single column sequences. See `test/test_union_all.py`
1.7.0rc.1
Remove Persistor's `query_by_keys()` and `df_from_query_by_keys()`. Use `query()` and `df_from_query()` with keyword argument `keys` and `parameters`.
Add caching to Persistor's `df_from_query()` with keyword argument `cache` as the filesystem location to store compressed picked dictionaries of sql -> DataFrame. Caching compares the literal rendered sql only: it is not aware of volatile functions called on the server (like current time) or side effects (like temp tables).
Add rendering of multi-column key values to Persistor's `query()` and `df_from_query()`. Users pass DataFrames instead of single column sequences. See `test/test_union_all.py`