Spark 3.4: Set metricsReporter for CopyOnWriteScan/MergeOnReadScan/IncrementalAppendScan/ChangelogScan

[wForget]

close #8444

test sql:

CREATE TABLE wangzhen_test_iceberg_20230824_t3 (id bigint, name string)
USING iceberg;

INSERT INTO wangzhen_test_iceberg_20230824_t3 VALUES (4, 'd'),(5, 'e'),(6, 'f');
delete from wangzhen_test_iceberg_20230824_t3 where id = 5;

ui before:

after:

[amogh-jahagirdar]

This is a good find @wForget , thanks for the fix. I'll wait for @aokolnychyi @RussellSpitzer and others to provide their inputs

[ConeyLiu]

Just curious, Is the after picture correct? Why a table with just three records can scan 41 manifests?

[wForget]

Just curious, Is the after picture correct? Why a table with just three records can scan 41 manifests?

I repeatedly executed INSERT and DELETE statement while debugging. I re-executed it using a new table and the result was correct.

[ConeyLiu]

+1

[aokolnychyi]

@wForget, this looks good to me. Can we follow up with a few tests, similar to what was done for regular scans?

[aokolnychyi]

Thanks, @wForget! Thanks for reviewing, @amogh-jahagirdar @ConeyLiu!

[wForget]

@wForget, this looks good to me. Can we follow up with a few tests, similar to what was done for regular scans?

@aokolnychyi Thanks, I will try to add relevant test cases.