Abhishek Rawat fa525dfdf7 IMPALA-7876: COMPUTE STATS TABLESAMPLE is not updating number of estimated rows ...

'COMPUTE STATS TABLESAMPLE' uses a child query with following function
'ROUND(COUNT(*) / <effective_sample_perc>)' for computing the row count.
The 'ROUND()' fn returns the row count as a DECIMAL type. The
'CatalogOpExecutor' (CatalogOpExecutor::SetTableStats) expects the row
count as a BIGINT type. Due to this data type mismatch the table stats
(Extrap #Rows) doesn't get set.

Adding an explicit CAST to BIGINT for the ROUND function results in the
table stats (Extrap #Rows) getting set properly.

Fixed both 'custom_cluster/test_stats_extrapolation.py' and
'metadata/test_stats_extrapolation.py' so that they can catch issues
like this, where table stats are not set when using
'COMPUTE STATS TABLESAMPLE'.

Testing:
- Ran core tests.

Change-Id: I88a0a777c2be9cc18b3ff293cf1c06fb499ca052
Reviewed-on: http://gerrit.cloudera.org:8080/16712
Reviewed-by: Tim Armstrong <tarmstrong@cloudera.com>
Tested-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>

2020-11-13 09:07:29 +00:00

3.2 KiB

Raw Blame History

View Raw