[SPARK-48188][SQL] Consistently use normalized plan for cache #46465
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
yaooqinn
approved these changes
May 8, 2024
|
Thank you @cloud-fan, merged to master |
JacobZheng0927
pushed a commit
to JacobZheng0927/spark
that referenced
this pull request
May 11, 2024
### What changes were proposed in this pull request? We must consistently use normalized plans for cache filling and lookup, or inconsistency will lead to cache misses. To guarantee this, this PR makes `CacheManager` the central place to do plan normalization, so that callers don't need to care about it. Now most APIs in `CacheManager` take either `Dataset` or `LogicalPlan`. For `Dataset`, we get the normalized plan directly. For `LogicalPlan`, we normalize it before further use. The caller side should pass `Dataset` when invoking `CacheManager`, if it already creates `Dataset`. This is to reduce the impact, as extra creation of `Dataset` may have perf issues or introduce unexpected analysis exception. ### Why are the changes needed? Avoid unnecessary cache misses for users who add custom normalization rules ### Does this PR introduce _any_ user-facing change? No, perf only ### How was this patch tested? existing tests ### Was this patch authored or co-authored using generative AI tooling? no Closes apache#46465 from cloud-fan/cache. Authored-by: Wenchen Fan <wenchen@databricks.com> Signed-off-by: Kent Yao <yao@apache.org>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
We must consistently use normalized plans for cache filling and lookup, or inconsistency will lead to cache misses.
To guarantee this, this PR makes
CacheManagerthe central place to do plan normalization, so that callers don't need to care about it. Now most APIs inCacheManagertake eitherDatasetorLogicalPlan. ForDataset, we get the normalized plan directly. ForLogicalPlan, we normalize it before further use.The caller side should pass
Datasetwhen invokingCacheManager, if it already createsDataset. This is to reduce the impact, as extra creation ofDatasetmay have perf issues or introduce unexpected analysis exception.Why are the changes needed?
Avoid unnecessary cache misses for users who add custom normalization rules
Does this PR introduce any user-facing change?
No, perf only
How was this patch tested?
existing tests
Was this patch authored or co-authored using generative AI tooling?
no