I think we should just return an ndarray from Column.to_pandas. Open a JIRA and we can change and make sure it's what we want
[ Full content available at: https://github.com/apache/arrow/pull/2535 ] This message was relayed via gitbox.apache.org for [email protected]
