Test which Arrow and Pandas datatypes are supported in write APIs

[wjones127]

Would be good to maintain unit tests that cover all the data types and understand roundtrip behavior.

Create two unit tests: one for PyArrow types and one for Pandas types. For each, create a table/DataFrame where there is one column for each possible data type. Verify the table can be roundtripped. If there are any that don't roundtrip, comment them out and create a new GitHub issue to track support for the data type.

[changhiskhan]

@wjones127 this could be a good first issue for europython sprint participants. When you get a chance would you mind adding more description here to limit the scope (e.g., which data types are still needed, how best to test these without writing a bunch of duplicate code etc) ?