lgbo-ustc commented on issue #6583: URL: https://github.com/apache/incubator-gluten/issues/6583#issuecomment-2255032811
It's a bug from https://github.com/ClickHouse/ClickHouse/pull/55809. Some values of a nullable column string are marked as nulls only with the null map, but the nested column values are not empty. For example, a nullable column string could be as following | row | null map |nested column string | | ----------- | ----------- | ----------- | | 0 | 0 | abcd | | 1 | 1 | aa | | 2 | | 0 | sdfsdf| | 3 | 1 | | | 4 |0 |sfsdf| `CH` doesn't require the nested value must be the default value. When `serializeValueIntoMemoryWithNull` is called on this case, it will allocate 11 bytes for row 1, but it just copy only one byte into the arena pool. later apply `deserializeAndInsertFromArena` will fail. Since it only move one byte on row 1, and cause following rows cannot be deserilaized. -- 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]
