MINOR: reuse pseudo-topic in FKJoin#8296
MINOR: reuse pseudo-topic in FKJoin#8296vvcephei merged 4 commits intoapache:trunkfrom vvcephei:minor-fix-pseudo-topic
Conversation
|
Hey @abbccdda , can you take a look at this minor patch? It's probably not a big deal, but we should use the same pseudo-topic for the same purpose. |
| //While we can use the source topic from where the events came from, we shouldn't serialize against it | ||
| //as it causes problems with the confluent schema registry, which requires each topic have only a single | ||
| //registered schema. | ||
| final String dummySerializationTopic = context().topic() + "-join-resolver"; |
There was a problem hiding this comment.
I just missed this last time I made a pass and cleaned up all the pseudo-topics. We should have passed in the common one for hashing LHS values, which is declared in KTableImpl (which is what we're doing now).
abbccdda
left a comment
There was a problem hiding this comment.
Could we add a unit test for this change?
| final SubscriptionResolverJoinProcessorSupplier<K, V, VO, VR> resolverProcessorSupplier = new SubscriptionResolverJoinProcessorSupplier<>( | ||
| primaryKeyValueGetter, | ||
| valSerde == null ? null : valSerde.serializer(), | ||
| valueHashSerdePseudoTopic, |
There was a problem hiding this comment.
Could we move the initialization of valueHashSerdePseudoTopic closer to where it is firstly used?
There was a problem hiding this comment.
This thought is reasonable, but I put all the pseudo-topic definitions together so that we could see what they all are at a glance.
There was a problem hiding this comment.
Plus, the first use is the very first processor in the FK join, so it couldn't move down by much anyway :)
|
Thanks for the review, @abbccdda ! Good thinking.. I've added the test. |
* apache-github/trunk: (39 commits) MINOR: cleanup and add tests to StateDirectoryTest (apache#8304) HOTFIX: StateDirectoryTest should use Set instead of List (apache#8305) MINOR: Fix build and JavaDoc warnings (apache#8291) MINOR: Fix kafka.server.RequestQuotaTest missing new ApiKeys. (apache#8302) KAFKA-9712: Catch and handle exception thrown by reflections scanner (apache#8289) KAFKA-9670; Reduce allocations in Metadata Response preparation (apache#8236) MINOR: fix Scala 2.13 build error introduced in apache#8083 (apache#8301) MINOR: enforce non-negative invariant for checkpointed offsets (apache#8297) MINOR: comment apikey types in generated switch (apache#8201) MINOR: Fix typo in CreateTopicsResponse.json (apache#8300) KIP-546: Implement describeClientQuotas and alterClientQuotas. (apache#8083) KAFKA-6647: Do note delete the lock file while holding the lock (apache#8267) KAFKA-9677: Fix consumer fetch with small consume bandwidth quotas (apache#8290) KAFKA-9533: Fix JavaDocs of KStream.transformValues (apache#8298) MINOR: reuse pseudo-topic in FKJoin (apache#8296) KAFKA-6145: Pt 2. Include offset sums in subscription (apache#8246) KAFKA-9714; Eliminate unused reference to IBP in `TransactionStateManager` (apache#8293) KAFKA-9718; Don't log passwords for AlterConfigs in request logs (apache#8294) KAFKA-8768: DeleteRecords request/response automated protocol (apache#7957) KAFKA-9685: Solve Set concatenation perf issue in AclAuthorizer ...
Reuse the same pseudo-topic for serializing the LHS value in the foreign-key join resolver as we originally used to serialize it before sending the subscription request. Reviewers: Boyang Chen <boyang@confluent.io>
Reuse the same pseudo-topic for serializing the LHS value in the foreign-key join resolver as we originally used to serialize it before sending the subscription request. Reviewers: Boyang Chen <boyang@confluent.io>
Reuse the same pseudo-topic for serializing the LHS value in the foreign-key join resolver as we originally used to serialize it before sending the subscription request. Reviewers: Boyang Chen <boyang@confluent.io>
Reuse the same pseudo-topic for serializing the LHS value in the foreign-key join resolver as
we originally used to serialize it before sending the subscription request.
Committer Checklist (excluded from commit message)