Dandandan commented on PR #23282: URL: https://github.com/apache/datafusion/pull/23282#issuecomment-4873611157
> > * To get coverage while not having to write/maintain a lot of new tests we could consider to run it against `sqllogictest` tests (starting with a subset probably to make it feasible): essentially running `sqllogictest` tests under the distributed execution mechanism. This would help to get coverage for more rare aggregation / window functions, join types, execution plan properties, etc. (that a smaller test like only running e.g. TPC-H would miss). > > Many of the `sqllogictest` tests only query a few rows, sometimes just evaluate literals. These may not be the best fit for distributed execution. I realize that, yet I think due to the size it will test much more than only a tpch / tpc-ds style benchmark test is going to do (I think trying to run the entire test _probably_ will uncover quite some issues and also avoid a wide range of issues). Having some extra tests (e.g. with more/larger files) to stress-test distributed usage would be great though. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
