[SPARK-51818][CONNECT] Move QueryExecution creation to AnalyzeHandler and don't Execute for AnalyzePlanRequests #50605
+70
−35
Conversation
This file contains hidden or 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
peterpashkin
changed the title
peterpashkin
changed the title
empty formatting test if that is good typo fix only explain change isLocal execution okay only Schema just work please do all correct include small fix squash
peterpashkin
force-pushed
the
peter-pashkin/MoveAnalyzeAndSkipExecution
branch
from
9a092f3 to
45bf4db
Compare
added 2 commits
April 16, 2025 16:06
Member
|
cc @vicennial @hvanhovell FYI |
vicennial
reviewed
Apr 23, 2025
Contributor
vicennial
left a comment
vicennial
left a comment
There was a problem hiding this comment.
Thanks for the changes!
The fact that analysis calls were executing commands looks like an unfortunate bug that slipped through.
Requests:
- Please fill in the "User Facing Changes" section. Include a simple example where the behaviour differs (IIUC,
spark.sql("<some command>").schemawould now have it's behaviour corrected) - Add a server-side test that verifies that the DF was in fact, not executed
Author
|
Thanks Akhil, will add the tests and User Facing Changes. Actually spark.sql(...).schema is still executing the command because commands with .sql() will get executed eagerly. This PR only fixes sending AnalyzePlanRequests not executing the request, like session.analyze(plan). |
vicennial
approved these changes
Apr 29, 2025
Contributor
vicennial
left a comment
vicennial
left a comment
There was a problem hiding this comment.
LGTM, thanks for the fix!
hvanhovell
reviewed
Apr 30, 2025
.../server/src/main/scala/org/apache/spark/sql/connect/service/SparkConnectAnalyzeHandler.scala
Show resolved
Hide resolved
hvanhovell
reviewed
Apr 30, 2025
.../server/src/main/scala/org/apache/spark/sql/connect/service/SparkConnectAnalyzeHandler.scala
Outdated
Show resolved
Hide resolved
HyukjinKwon
approved these changes
May 7, 2025
Member
|
Merged to master. |
yhuang-db
pushed a commit
to yhuang-db/spark
that referenced
this pull request
Jun 9, 2025
… and don't Execute for AnalyzePlanRequests ### What changes were proposed in this pull request? Analyze Plan Requests for Schema should not trigger an Execute on the Logical Plan, currently when sending an AnalyzePlanRequest with a command that gets executed eagerly the Dataset.ofRows(logicalPlan) call executes the underlying command. We do not want this to happen when doing AnalyzePlan. So instead we construct the LogicalPlan with the CommandExecutionMode.SKIP and return the resulting schema that way. https://issues.apache.org/jira/browse/SPARK-51818 ### Why are the changes needed? SQL commands that get sent via an AnalyzePlanRequest get executed eagerly right now, this PR fixes that ### Does this PR introduce _any_ user-facing change? When calling .schema on DataFrame via Spark Connect the plan saved in the DataFrame is not executed anymore, that was the case beforehand. Example: spark.newDataFrame(plan: proto.Plan).schema with plan encoding some SQL command that gets executed eagerly like DROP TABLE the current behavior would execute the SQL command. This will not happen anymore after this change. ### How was this patch tested? Added Test for sending an AnalyzePlanRequest with Drop Table and making sure the table was not dropped ### Was this patch authored or co-authored using generative AI tooling? No Closes apache#50605 from peterpashkin/peter-pashkin/MoveAnalyzeAndSkipExecution. Authored-by: Peter Pashkin <peter.pashkin@databricks.com> Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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?
Analyze Plan Requests for Schema should not trigger an Execute on the Logical Plan, currently when sending an AnalyzePlanRequest with a command that gets executed eagerly the Dataset.ofRows(logicalPlan) call executes the underlying command. We do not want this to happen when doing AnalyzePlan. So instead we construct the LogicalPlan with the CommandExecutionMode.SKIP and return the resulting schema that way.
https://issues.apache.org/jira/browse/SPARK-51818
Why are the changes needed?
SQL commands that get sent via an AnalyzePlanRequest get executed eagerly right now, this PR fixes that
Does this PR introduce any user-facing change?
When calling .schema on DataFrame via Spark Connect the plan saved in the DataFrame is not executed anymore, that was the case beforehand. Example: spark.newDataFrame(plan: proto.Plan).schema with plan encoding some SQL command that gets executed eagerly like DROP TABLE the current behavior would execute the SQL command. This will not happen anymore after this change.
How was this patch tested?
Added Test for sending an AnalyzePlanRequest with Drop Table and making sure the table was not dropped
Was this patch authored or co-authored using generative AI tooling?
No