msck 是否修复触发表统计信息生成

Question

i would like to know whether running msck repair <table> on table will generate table statistics for CBO when:

hive.cbo.enable=true

hive.stats.autogather=true

or do i have to use analyze <tablename> compute statistics .

Thanks

Answer 1

It can gather stats when running REPAIR. But it is considered as a bug because the REPAIR command runs unexpectedly slow in this case.

There are a number of related JIRAs:

HIVE-18743 HIVE-19489 HIVE-17478 SPARK-17063

hive.stats.autogather=false before REPAIR is a workaround.

Since it is considered like a bug, you better do not rely on this feature and do analyze when necessary.

Found this related question: https://stackoverflow.com/a/57885181/2700344