SEZ9 commented on issue #13330:
URL: https://github.com/apache/seatunnel/issues/13330#issuecomment-7050601206
@SEPURI-SAI-KRISHNA thanks for checking each point. Your comment answers all
three, and I have no open asks left on them.
- **Catch-all vs. ranges:** The reasoning holds as you describe it. A range
only matches its own type bracket, and the catch-all only takes null or
missing, NaN, and the types no range covers, so non-array values land in
exactly one split. The IT data without `c_key`, with each `c_id` written
exactly once, is the right evidence for `{key: null}` matching a missing key.
Leaving arrays as the exception named in the docs tip is fine.
- **`match.query` on the split key:** `testEverySplitKeepsTheMatchQuery` is
the case I wanted, since it splits by `_id` with `{_id: {$gte: 1500}}` and
asserts the `$and` in every split.
- **IT data:** The NaN values of `c_key` and the job split by `c_nested.k`
close the gaps. As you note, that job still reads each document once without
the dotted fix (1 split instead of 18), so the IT alone would not catch a
regression there. The unit test for the value read along the dotted path is
what guards it, and you say it fails on the commit before, so that is covered.
The three-commit layout (1+2, 3, 4) is easy to review, so there is no need
to split it into separate PRs. Any remaining detail can be handled in the PR
review.
<!-- streview-comment:2478 -->
--
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]