alex-balikov opened a new pull request, #38405:
URL: https://github.com/apache/spark/pull/38405
<!--
Thanks for sending a pull request! Here are some tips for you:
1. If this is your first time, please read our contributor guidelines:
https://spark.apache.org/contributing.html
2. Ensure you have added or run the appropriate tests for your PR:
https://spark.apache.org/developer-tools.html
3. If the PR is unfinished, add '[WIP]' in your PR title, e.g.,
'[WIP][SPARK-XXXX] Your PR title ...'.
4. Be sure to keep the PR description updated to reflect all changes.
5. Please write your PR title to summarize what this PR proposes.
6. If possible, provide a concise example to reproduce the issue for a
faster review.
7. If you want to add a new configuration, please read the guideline first
for naming configurations in
'core/src/main/scala/org/apache/spark/internal/config/ConfigEntry.scala'.
8. If you want to add or modify an error type or message, please read the
guideline first in
'core/src/main/resources/error/README.md'.
-->
### What changes were proposed in this pull request?
<!--
Please clarify what changes you are proposing. The purpose of this section
is to outline the changes and how this PR fixes the issue.
If possible, please consider writing useful notes for better and faster
reviews in your PR. See the examples below.
1. If you refactor some codes with changing classes, showing the class
hierarchy will help reviewers.
2. If you fix some SQL features, you can provide some references of other
DBMSes.
3. If there is design documentation, please add the link.
4. If there is a discussion in the mailing list, please add the link.
-->
This PR fixes the input late record filtering done by stateful operators to
allow for chaining of stateful operators. Currently stateful operators are
initialized with the current microbatch watermark and perform both input late
record filtering and state eviction (e.g. producing aggregations) using the
same watermark value. The state evicted (or aggregates produced) due to
watermark advancing is behind the watermark and thus effectively late - if a
following stateful operator consumes the output of the previous one, the input
records will be filtered as late.
This PR provides two watermark values to the stateful operators - one from
the previous microbatch to be used for late record filtering and the one from
the current microbatch (as in the existing code) to be used for state eviction.
This solves the above problem of the broken late record filtering.
Note that this PR still does not solve the issue of time-interval stream
join producing records delayed against the watermark. Therefore time-interval
streaming join followed by stateful operators is still not supported. That will
be fixed in a follow up PR (and a SPIP) effectively replacing the single global
watermark with conceptually watermarks per operator.
Also, the stateful operator chains unblocked by this PR (e.g. a chain of
window aggregations) are still blocked by the unsupported operations checker.
The new test for these scenarios - MultiStatefulOperatorsSuite has to
explicitly disable the unsupported ops check. This again will be fixed in a
follow-up PR.
### Why are the changes needed?
<!--
Please clarify why the changes are needed. For instance,
1. If you propose a new API, clarify the use case for a new API.
2. If you fix a bug, you can clarify why it is a bug.
-->
The PR allows Spark Structured Streaming to support chaining of stateful
operators e.g. chaining of time window aggregations which is a meaningful
streaming scenario.
### Does this PR introduce _any_ user-facing change?
<!--
Note that it means *any* user-facing change including all aspects such as
the documentation fix.
If yes, please clarify the previous behavior and the change this PR proposes
- provide the console output, description and/or an example to show the
behavior difference if possible.
If possible, please also clarify if this is a user-facing change compared to
the released Spark versions or within the unreleased branches such as master.
If no, write 'No'.
-->
With this PR, chains of stateful operators will be supported in Spark
Structured Streaming.
### How was this patch tested?
<!--
If tests were added, say they were added here. Please make sure to add some
test cases that check the changes thoroughly including negative and positive
cases if possible.
If it was tested in a way different from regular unit tests, please clarify
how you tested step by step, ideally copy and paste-able, so that other
reviewers can test and check, and descendants can verify in the future.
If tests were not added, please describe why they were not added and/or why
it was difficult to add.
If benchmark tests were added, please run the benchmarks in GitHub Actions
for the consistent environment, and the instructions could accord to:
https://spark.apache.org/developer-tools.html#github-workflow-benchmarks.
-->
Added a new test suite - MultiStatefulOperatorsSuite
--
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]
