mopcup opened a new pull request, #13934: URL: https://github.com/apache/arrow/pull/13934
I have experimentally implemented .keep_all = TRUE using hash_one. This implementation will not pass the test skipped currently. There are two known issues. First, since hash_one is currently executed for each field, data may be corrupted if fields contain nulls (i.e., nulls are skipped and changed to other values, this may generate rows that did not exist in raw data). It will be a serious problem when group vars can not uniquely determine the first row. `.keep_all = TRUE` will make a warning about this, but I'm concerned that the user may miss it. Second, it does not support the factor type. This may be solved when FunctionOption for hash_one is implemented. ScalarAggregateOptions is currently used. -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
