coderfender commented on PR #6938:
URL:
https://github.com/apache/datafusion-comet/pull/6938#issuecomment-6623990571
Thanks @andygrove and @comphead. Addressed in 7c2c956af:
- **Row order (@andygrove):** you were right, and it starts earlier than
8192 rows. A test comparing rows in order showed the native output is
right-batch-major, so the order differs as soon as the right side arrives in
more than one batch, even one per Parquet file. Spark reports no output
ordering for `CartesianProductExec`, so I noted it in `operators.md` ("Row
order can differ from Spark's").
- **Config note (@andygrove):** it now links apache/datafusion#26383.
- **Types (@comphead):** `cartesian_product.sql` has a `cp_types` left side
with string, decimal, timestamp, date, binary, boolean, double (NaN, `-0.0` and
`0.0`), array, struct and map columns, with top-level and nested NULLs and
empty values. It runs natively and matches Spark with AQE on and off.
- **Description (@comphead):** updated. It now says "Part of #199", with
multi-partition inputs and spilling tracked in #7476.
--
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]
