[
https://issues.apache.org/jira/browse/IMPALA-13636?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17910426#comment-17910426
]
ASF subversion and git services commented on IMPALA-13636:
----------------------------------------------------------
Commit f1de0c392f4c5fe1df653956595a6f6699dd100e in impala's branch
refs/heads/master from Riza Suminto
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=f1de0c392 ]
IMPALA-13636: Fix target file_format in TestTpcdsInsert
TestTpcdsInsert creates a temporary table to test insert functionality.
It has three problems:
1. It does not use unique_database parameter, so the temporary table is
not cleaned up after test finished.
2. It ignores file_format from test vector, causing inconsistency in the
temporary table's file format. str_insert is always in PARQUET format,
while store_sales_insert is always in TEXTFILE format.
3. text file_format dimension is never exercised, because
--workload_exploration_strategy in run-all-tests.sh does not
explicitly list tpcds-insert workload.
This patch fixes all three problems and few flake8 warnings in
test_tpcds_queries.py.
Testing:
- Run bin/run-all-tests.sh with
EXPLORATION_STRATEGY=exhaustive
EE_TEST=true
EE_TEST_FILES="query_test/test_tpcds_queries.py::TestTpcdsInsert"
Verified that the temporary table format follows file_format
dimension.
Change-Id: Iea621ec1d6a53eba9558b0daa3a4cc97fbcc67ae
Reviewed-on: http://gerrit.cloudera.org:8080/22291
Reviewed-by: Michael Smith <[email protected]>
Reviewed-by: Csaba Ringhofer <[email protected]>
Tested-by: Riza Suminto <[email protected]>
> TestTpcdsInsert ignores file_format from test vector
> ----------------------------------------------------
>
> Key: IMPALA-13636
> URL: https://issues.apache.org/jira/browse/IMPALA-13636
> Project: IMPALA
> Issue Type: Improvement
> Components: Test
> Affects Versions: Impala 4.4.0
> Reporter: Riza Suminto
> Assignee: Riza Suminto
> Priority: Minor
>
> TestTpcdsInsert creates a temporary table to test insert functionality. It
> has three problems:
> # It does not use unique_database parameter, so the temporary table is not
> cleaned up after test finished.
> # It ignores file_format from test vector, causing inconsistency in the
> temporary table's file format.
> [str_insert|https://github.com/apache/impala/blob/14b959d7bf7104d0172ce41df909fc0835067644/testdata/workloads/tpcds-insert/queries/expr-insert.test#L6]
> is always in PARQUET format, while
> [store_sales_insert|https://github.com/apache/impala/blob/14b959d7bf7104d0172ce41df909fc0835067644/testdata/workloads/tpcds-insert/queries/partitioned-insert.test#L6]
> is always in TEXTFILE format.
> # Insert in TEXTFILE format is never exercised, even in exhaustive
> exploration.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]