[SPARK-50903][CONNECT] Cache logical plans after analysis #49584
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
changgyoopark-db
force-pushed
the
SPARK-50903
branch
2 times, most recently
from
db8db47 to
8b1df40
Compare
changgyoopark-db
changed the title
changgyoopark-db
changed the title
changgyoopark-db
changed the title
changgyoopark-db
changed the title
changgyoopark-db
force-pushed
the
SPARK-50903
branch
2 times, most recently
from
6fc5c1b to
38d2d1d
Compare
changgyoopark-db
changed the title
changgyoopark-db
force-pushed
the
SPARK-50903
branch
5 times, most recently
from
022f1e3 to
4264bdd
Compare
changgyoopark-db
force-pushed
the
SPARK-50903
branch
4 times, most recently
from
e948a7c to
5962637
Compare
changgyoopark-db
force-pushed
the
SPARK-50903
branch
from
5962637 to
337871f
Compare
changgyoopark-db
force-pushed
the
SPARK-50903
branch
from
337871f to
9748999
Compare
changgyoopark-db
changed the title
changgyoopark-db
force-pushed
the
SPARK-50903
branch
2 times, most recently
from
7b7430f to
622bac9
Compare
changgyoopark-db
force-pushed
the
SPARK-50903
branch
from
622bac9 to
18c2344
Compare
changgyoopark-db
force-pushed
the
SPARK-50903
branch
5 times, most recently
from
ca61632 to
1f8594e
Compare
|
@xi-db FYI
|
changgyoopark-db
force-pushed
the
SPARK-50903
branch
from
1f8594e to
4aac2c3
Compare
changgyoopark-db
changed the title
changgyoopark-db
force-pushed
the
SPARK-50903
branch
2 times, most recently
from
43b810f to
5981821
Compare
hvanhovell
reviewed
Jan 30, 2025
sql/connect/server/src/main/scala/org/apache/spark/sql/connect/service/SessionHolder.scala
Outdated
Show resolved
Hide resolved
hvanhovell
reviewed
Jan 30, 2025
sql/connect/server/src/main/scala/org/apache/spark/sql/connect/service/SessionHolder.scala
Show resolved
Hide resolved
hvanhovell
reviewed
Jan 30, 2025
| logDebug(s"Cache a lazyily analyzed logical plan for '$rel': $plan") | ||
| planCache.get.put(rel, plan) | ||
| } else { | ||
| val plan = df.queryExecution.analyzed |
There was a problem hiding this comment.
We may have to add some invalidation logic here. The problem is that some of objects (tables/views/udfs) used in the query can change, in that case we may want to validate that the objects used are the most recent ones.
There was a problem hiding this comment.
Agreed. I'll need to think about the interface.
There was a problem hiding this comment.
Added LogicalPlan.isOutdated and updated the usePlanCache method.
-> Calling 'refresh' seemed suboptimal, as schema changes often require a re-planning of the plan.
changgyoopark-db
force-pushed
the
SPARK-50903
branch
from
259022b to
ca3cebf
Compare
changgyoopark-db
force-pushed
the
SPARK-50903
branch
from
ed80473 to
1ce371e
Compare
changgyoopark-db
force-pushed
the
SPARK-50903
branch
from
1ce371e to
649ba6c
Compare
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?
Update the plan cache after logical plans are analyzed.
Additionally, this PR includes,
Why are the changes needed?
The session local plan cache may store unanalyzed logical plans, which causes them to be repeatedly analyzed, resulting in performance degradation.
To be specific, the session local plan cache usage pattern is,
Where an unanalyzed plan may be cached during transformation, and Dataset.ofRows analyzes the unanalyzed plan. This leads to a situation where the same query may hit the cache, returning an unanalyzed plan, and Dataset.ofRows again analyzes the same plan.
Does this PR introduce any user-facing change?
No.
How was this patch tested?
testOnly org.apache.spark.sql.connect.service.SparkConnectSessionHolderSuite
Was this patch authored or co-authored using generative AI tooling?
No.