Because each DBQL table has its own separate cache, the system may fill up the cache for one table and log to that DBQL table before it logs into another DBQL table. This results in slightly different CollectTimeStamp values. Therefore, do not include CollectTimeStamp fields in joins.
If you need to join different DBQL tables, use the QueryID field. If you plan on comparing values between different DBQL tables, you can flush the DBQL caches to the dictionary tables. See Options for Flushing the DBQL Cache.
Teradata recommends that you ignore the orphan rows and do not copy them over when you copy the DBQL tables out of DBC and into temporary or history tables in another database.