bamaer commented on code in PR #9480:
URL: https://github.com/apache/hop/pull/9480#discussion_r4573552228


##########
plugins/transforms/joinrows/src/main/java/org/apache/hop/pipeline/transforms/joinrows/JoinRowsMeta.java:
##########
@@ -321,6 +321,15 @@ public boolean excludeFromRowLayoutVerification() {
     return true;
   }
 
+  /**
+   * A cartesian product needs every row of every input in one place. With 
several copies each copy
+   * only receives a part of every stream, so combinations would go missing.
+   */
+  @Override
+  public boolean supportsMultiCopyExecution() {

Review Comment:
   **Should fix:** several copies give the full result when the other streams 
copy their rows instead of distributing them. Join Rows ×2 with 100 main rows 
and 10 other rows: Distribute gives 500 rows, while setting the "other" 
transform to Copy rows gives all 1000. Returning `false` here removes "Number 
of copies" from the GUI for that setup. Pipelines that already have copies > 1 
also can't be set back to 1 copy in the GUI any more. Suggest dropping this 
override and `doesNotSupportMultipleCopies`, or moving it to its own issue.



##########
docs/hop-user-manual/modules/ROOT/pages/pipeline/transforms/joinrows.adoc:
##########
@@ -47,6 +47,10 @@ To build a cartesian product on these engines, add the same 
constant field to bo
 See 
xref:pipeline/beam/getting-started-with-beam.adoc#_unsupported_transforms[Getting
 started with Beam] and 
xref:pipeline/pipeline-run-configurations/native-spark-pipeline-engine.adoc#transforms-not-supported-on-native-spark[Transforms
 not supported on Native Spark].
 ====
 
+NOTE: Run this transform in a single copy.

Review Comment:
   Suggest instead: `NOTE: With several copies of this transform, set the 
transforms of the other streams to Copy rows. Otherwise each copy only receives 
a part of those streams and combinations go missing.`



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to