alamb opened a new pull request #8400: URL: https://github.com/apache/arrow/pull/8400
This is a proposed approach - if people like this approach I will write proper tests (to ensure that the cast kernel support and this function remain in sync). This PR brings DataFusion to parity with the type casting supported by arrow and allows DataFusion to plan all casts that are supported by the arrow cast kernel Previously the notions of coercion and casting were somewhat conflated. I have tried to clarify them in https://github.com/apache/arrow/pull/8399 and this PR. See also https://github.com/apache/arrow/pull/8340#discussion_r501257096 for more discussion. I personally want this functionality so when I add support for `DictionaryArray` casts in Arrow (https://github.com/apache/arrow/pull/8346) they can also be used in DataFusion. Codewise, I am concerned about the duplication in logic between this and cast.rs. However, the test I have in mind will ensure they don't get out of sync Questions for reviewers: 1) Would this code / test be better to include in the arrow cast kernel (cast.rs) itself? To keep the "support" matrix code closer to the implementation? ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
