comphead commented on PR #26266: URL: https://github.com/apache/datafusion/pull/26266#issuecomment-6297302463
Thanks, pushed `fb1b29b9c`. - The generated rows check now runs before the all-scalar shortcut, with the real batch size. A unit test calls `invoke_with_args` with `n = 2^20` and 8192 rows. The folded literal case still gets through, as you noted. - Nullability is matched exactly again, as in Spark. The slt case now expects the plan-time error for `List(non-null Int64)` followed by `[2, NULL]`. - Only `Int64` literals are widened in an `INT` column, so `TINYINT` and `SMALLINT` literals are rejected as in Spark. A unit test covers it. - The limit error now reads "generating 2000000000 rows for each of 1 input rows exceeds the limit of 16777216 rows per batch". - The limit and the `n > 2048` consequence for 8192-row batches are in the user-facing changes section of the PR description. -- 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]
