Zoltan Borok-Nagy has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/21423 )

Change subject: IMPALA-12732: Add support for MERGE statements for Iceberg 
tables
......................................................................


Patch Set 13: Code-Review+1

(9 comments)

Left a few comments, but I think we are close to the finish line :)

http://gerrit.cloudera.org:8080/#/c/21423/11/be/src/exec/iceberg-merge-node.h
File be/src/exec/iceberg-merge-node.h:

http://gerrit.cloudera.org:8080/#/c/21423/11/be/src/exec/iceberg-merge-node.h@110
PS11, Line 110: ergeCase*> matched_cases_;
> It would require the null checks to be moved into EvaluateCases, as the res
I meant we could just save 'Tuple* previous_row_target_tuple' instead of 
'last_row_'. And IsDuplicateRow() could still have 'TupleRow* actual_row' as 
its parameter. Duplicate check could be a bit faster and simpler, but feel free 
to ignore this comment if you think it doesn't make too much sense.


http://gerrit.cloudera.org:8080/#/c/21423/11/be/src/exec/iceberg-merge-node.h@134
PS11, Line 134:
> The other way to initialize non-trivial object is a separate line like 'con
Ack


http://gerrit.cloudera.org:8080/#/c/21423/13/be/src/exec/iceberg-merge-node.cc
File be/src/exec/iceberg-merge-node.cc:

http://gerrit.cloudera.org:8080/#/c/21423/13/be/src/exec/iceberg-merge-node.cc@251
PS13, Line 251: int target_tuple_idx
Is 'target_tuple_idx' needed? It's always target_tuple_idx_.


http://gerrit.cloudera.org:8080/#/c/21423/13/be/src/exec/iceberg-merge-node.cc@251
PS13, Line 251: TupleRow* previous_row
Is 'previous_row' needed? It's always last_row_.


http://gerrit.cloudera.org:8080/#/c/21423/13/fe/src/main/java/org/apache/impala/analysis/DmlStatementBase.java
File fe/src/main/java/org/apache/impala/analysis/DmlStatementBase.java:

http://gerrit.cloudera.org:8080/#/c/21423/13/fe/src/main/java/org/apache/impala/analysis/DmlStatementBase.java@134
PS13, Line 134:   }
nit: missing empty line after this


http://gerrit.cloudera.org:8080/#/c/21423/11/fe/src/main/java/org/apache/impala/analysis/IcebergMergeImpl.java
File fe/src/main/java/org/apache/impala/analysis/IcebergMergeImpl.java:

http://gerrit.cloudera.org:8080/#/c/21423/11/fe/src/main/java/org/apache/impala/analysis/IcebergMergeImpl.java@258
PS11, Line 258:
> Unfortunately, it's required, if it's a simple slash (/), then in terminate
I see, since this SQL statement generates a parse error anyway, you could just 
write "* /", or use /// instead of /** .. */ for the comments.


http://gerrit.cloudera.org:8080/#/c/21423/13/fe/src/main/java/org/apache/impala/planner/Planner.java
File fe/src/main/java/org/apache/impala/planner/Planner.java:

http://gerrit.cloudera.org:8080/#/c/21423/13/fe/src/main/java/org/apache/impala/planner/Planner.java@142
PS13, Line 142: {
nit: missing space


http://gerrit.cloudera.org:8080/#/c/21423/13/testdata/workloads/functional-query/queries/QueryTest/iceberg-merge-long.test
File 
testdata/workloads/functional-query/queries/QueryTest/iceberg-merge-long.test:

http://gerrit.cloudera.org:8080/#/c/21423/13/testdata/workloads/functional-query/queries/QueryTest/iceberg-merge-long.test@535
PS13, Line 535: row_regex:.*[0-9]+
Would it be possible to check actual results? I see currently you are using 
random, but maybe we could switch to dummy data.


http://gerrit.cloudera.org:8080/#/c/21423/13/tests/stress/test_merge_stress.py
File tests/stress/test_merge_stress.py:

http://gerrit.cloudera.org:8080/#/c/21423/13/tests/stress/test_merge_stress.py@100
PS13, Line 100: new_total
new_total seems unnecessary



--
To view, visit http://gerrit.cloudera.org:8080/21423
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I3416a79740eddc446c87f72bf1a85ed3f71af268
Gerrit-Change-Number: 21423
Gerrit-PatchSet: 13
Gerrit-Owner: Peter Rozsa <[email protected]>
Gerrit-Reviewer: Daniel Becker <[email protected]>
Gerrit-Reviewer: Gabor Kaszab <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Noemi Pap-Takacs <[email protected]>
Gerrit-Reviewer: Peter Rozsa <[email protected]>
Gerrit-Reviewer: Zoltan Borok-Nagy <[email protected]>
Gerrit-Comment-Date: Fri, 16 Aug 2024 10:31:48 +0000
Gerrit-HasComments: Yes

Reply via email to