lirui-apache commented on issue #4997: URL: https://github.com/apache/iceberg/issues/4997#issuecomment-1649555725
Hey guys, we're facing a similar issue where our table is also partitioned by `dt string`, and our user sometimes writes filters like `dt = 20230725` by mistake. The filter is translated into `cast(dt as int) = 20230725` and can't be pushed down to iceberg and the query ends up as a full table scan. I wonder whether there's a solution for this. We're considering disallowing full table scan in `SparkBatchQueryScan`, but that seems unfair because there's indeed a partition filter in the query. Trino handles such cases by evaluating the filter expression on split level, because the partition column is a known literal for each split. So perhaps we can do the same in spark? -- 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]
