One of our SQL instances has a large MDS database, and virtually 100% of the size is due to a 300+ million row table named
stg.Stg_Project_Leaf. It recently started blowing up our weekly index rebuilds, so I need to figure this out.
From what I can tell, this is a Leaf Member Staging Table for a "Project" entity, set up and automated (by a prior consultant) as a part of an ETL process that builds BI cubes/reports.
Looks like it has been processing about 10k rows per half hour for the past several years and has never been purged. When I check the
ImportStatus_ID, though, I see 0 rows waiting to be processed. (Most are 1 - succeeded, less than 1% are 2 - failed.)
Can I just
TRUNCATE this table?
This system is being used in an automated fashion, its not like people are logging into the MDS console and individually monitoring and rolling back changes. I found some references to stored procedures that purge these tables (
mdm.udpStagingClear or maybe
mdm.udpDeletedMembersPurge), but I'm not sure which of those is the one I'd need, nor do I know if they would blow up the log if they tried to do 300m rows at once.