Archives of the TeradataForum

Message Posted: Thu, 17 Nov 2005 @ 17:53:25 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

With V2R5, the collect timestamp for the LOG table is written when the cache is allocated, but rows aren't put there until the query completes (this is necessary for end times, etc).

However, for the other tables the collect timestamp is taken when the cache is flushed. In addition the rows are written to the OBJ and SQL tables at the start of the query. The step table is writes the rows to cache as each step completes.

Thanks mainly to Carrie for helping me get to this point, I hope I have passed on the info correctly!

Utilising the start time in the LOG and STEP tables, these can all be joined provided the cache is flushed regularly and you have an understanding of how long a query can run for.

Using the Long Query ID (DBS control setting) removes this complexity.

Regards

Keith.

	<Prev	Next>		<<First	<Prev	Next>	Last>>