icefury71 commented on issue #4230: NULL value support for all data types URL: https://github.com/apache/incubator-pinot/issues/4230#issuecomment-511595289 Updating this thread with some of the offline discussions I had with a few folks (Devesh and @fx19880617 ). This proposal is to filter out null values in the form of predicates (leaf nodes of query execution). This has 2 potential issues: 1) Incorrect results when transforms are used in aggregation functions (eg: MAX(ADD(A, B)) if either A or B is null, ADD should return NULL. However if we filter out nulls very early, then we will return a non null value here - which is not consistent with ANSI SQL. 2) Multiple aggregation functions in the same query: This also might be inconsistent if we filter out the nulls very early in query execution. Looks like a better approach might be to "tag" DocIDs as null or non-nulls and let the aggregation function (or another intermediate node) process it in the right manner. However, this will affect the entire query execution flow and is a much bigger effort than what's proposed in this issue. Thoughts ?
---------------------------------------------------------------- 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] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
