gaodayue opened a new issue #2668: [Compaction] Tablet of a static dataset can't be fully compacted into one rowset URL: https://github.com/apache/incubator-doris/issues/2668 **Describe the bug** I loaded one static dataset into a Doris table. I expect all tablets to be compacted into one rowset after some time since there is no more new data. But it turns out all tablets end up having 2 or 3 versions. ``` $ curl 'http://$be:$port/api/compaction/show?tablet_id=12492&schema_hash=1862098505' { "cumulative point": 229, "last cumulative failure time": "2020-01-06 16:39:05.724", "last base failure time": "2020-01-06 14:45:37.679", "last cumulative success time": "2020-01-02 18:37:35.086", "last base success time": "2020-01-02 19:01:21.834", "versions": [ "[0-228] 6 ", "[229-229] 1 ", "[230-230] 1 " ] } $ curl 'http://$be:$port/api/compaction/show?tablet_id=12508&schema_hash=1862098505' { "cumulative point": 230, "last cumulative failure time": "2020-01-06 17:56:42.175", "last base failure time": "2020-01-06 17:56:45.007", "last cumulative success time": "2020-01-02 19:24:21.047", "last base success time": "2020-01-02 19:32:26.552", "versions": [ "[0-229] 12 ", "[230-230] 2 " ] } ``` The reason I want all tablets to be base-compacted into one version is that the aggregated table has a long varchar key column which cause the merge_iterator to be slow, union_iterator can be utilized to boost query performance when there is only one version. **Environment** Latest master branch with default configurations.
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
