MINOR: cleanup KStream JavaDocs (5/N) - stream-globalTable-inner-join #18747
Conversation
streams/src/main/java/org/apache/kafka/streams/kstream/KStream.java
Outdated
Show resolved
Hide resolved
| @@ -1043,33 +1043,96 @@ public <VO, VR> KStream<K, VR> leftJoin(final KTable<K, VO> table, | |||
| } | |||
| } | |||
|
|
|||
| @SuppressWarnings("unchecked") | |||
| private <VO, VR> KStream<K, VR> doStreamTableJoin(final KTable<K, VO> table, | |||
There was a problem hiding this comment.
Just move this method from further below up here, where it belongs.
| * If a {@code KStream} input value is {@code null} the record will not be included in the join operation | ||
| * and thus no output record will be added to the resulting {@code KStream}. | ||
| * The key of the result record is the same as the stream record's key. | ||
| * If you need read access to the {@code KStream} key, use {@link #join(GlobalKTable, KeyValueMapper, ValueJoinerWithKey)}. |
There was a problem hiding this comment.
I just describe what the code currently does. But it seems it's not correct. We should not pass in the KStream record key, but the join key (ie, whatever keySelector returns).
The original KIP says (https://cwiki.apache.org/confluence/display/KAFKA/KIP-149%3A+Enabling+key+access+in+ValueTransformer%2C+ValueMapper%2C+and+ValueJoiner)
it seems like extending the interface to pass the join key along as well would be helpful
The question now is, can we change this directly? Even if it would be a bug-fix, is could be considered a "breaking change" (should we try to sneak a fix into 4.0 which would allows us to ship a breaking change)?
mjsax
force-pushed
the
minor-update-kstreams-javadocs-5
branch
from
103e097 to
4e504ef
Compare
No description provided.