github-actions[bot] commented on code in PR #69132:
URL: https://github.com/apache/doris/pull/69132#discussion_r4281737920
##########
regression-test/suites/external_table_p2/maxcompute/test_max_compute_timestamp.groovy:
##########
@@ -216,56 +186,6 @@ suite("test_max_compute_timestamp", "p2,external") {
sql """ use ${mc_db}"""
- sql """ set time_zone = "Asia/Shanghai" """
- qt_7_1 """ select * from datetime_tb1;"""
- qt_7_2 """ select * from datetime_tb1 where col1 > "2023-02-02
00:00:00.000";"""
- qt_7_3 """ select * from datetime_tb1 where col1 >= "2023-02-02
00:00:00.000";"""
- qt_7_4 """ select * from datetime_tb1 where col1 = "2023-02-02
00:00:00.000";"""
- qt_7_5 """ select * from datetime_tb1 where col1 <= "2023-02-02
00:00:00.000";"""
- qt_7_6 """ select * from datetime_tb1 where col1 < "2023-02-02
00:00:00.000";"""
- qt_7_7 """ select * from datetime_tb1 where col1 != "2023-02-02
00:00:00.000";"""
-
-
- qt_8_1 """ select * from timestamp_tb2;"""
- qt_8_2 """ select * from timestamp_tb2 where col1 > "2023-02-02
00:00:00.123456";"""
- qt_8_3 """ select * from timestamp_tb2 where col1 >= "2023-02-02
00:00:00.123456";"""
- qt_8_4 """ select * from timestamp_tb2 where col1 = "2023-02-02
00:00:00.123456";"""
Review Comment:
[P2] Retain an enabled TIMESTAMP predicate smoke test
Deleting the entire query block after switching to `${mc_catalog_name}_2`
leaves every remaining timestamp query on the catalog with
`mc.datetime_predicate_push_down=false`. The all-type suite still covers
enabled DATETIME/TIMESTAMP_NTZ, but has no TIMESTAMP column, so it cannot
exercise the separate TIMESTAMPTZ-to-ODPS conversion (including avoiding a
second timezone conversion). A regression there could discard matching rows
while all retained checks pass. Please keep a small TIMESTAMP equality/range
check on the enabled catalog in Shanghai and UTC; the full matrix is
unnecessary.
##########
regression-test/suites/external_table_p2/maxcompute/test_max_compute_partition_prune.groovy:
##########
@@ -117,213 +113,153 @@ suite("test_max_compute_partition_prune",
"p2,external") {
String mc_catalog_name = "test_max_compute_partition_prune"
- for (String enable_profile : ["true","false"] ) {
- sql """set enable_profile = ${enable_profile} """;
- sql """set fetch_splits_max_wait_time_ms = 10000 """;
-
- for (String num_partitions : ["1","10","100"] ) {
Review Comment:
[P2] Keep one query that forces partition batch scanning
Removing every `num_partitions_in_batch_mode` override makes this suite use
the default threshold of 1024, while its tables have only 2, 4, and 10
partitions. Consequently none of these queries exercises
`PluginDrivenScanNode.startSplit`, including its asynchronous partition
slicing, split queue, and completion handling. Other MaxCompute suites do not
lower the threshold, and the batch-mode unit tests cover the decision gate
rather than this pump. Please retain one multi-partition result check with
threshold 1 alongside the default pass; this preserves the distinct execution
path without restoring the expensive matrix.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
