Archives of the TeradataForum

Message Posted: Wed, 05 Apr 2006 @ 09:31:07 GMT

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

Hi Alan,

Just a little more detail on long, or "unique" query ID.

Unique query ID is a change in the way QueryID column gets populated to improve uniqueness. It addresses the problem of QueryID numbers getting reset in each PE upon restart.

The feature uses PE startup time to seed the upper digits of ProcID and QueryID, forcing uniqueness. The PE 9-digit start time is broken into two parts. The high portion added in front of ProcID, the low in front of QueryID. This PE start time is actually an accumulation of 10ths of a second since 1970 (or something similar to that), so the field is always getting bigger, always slightly different.

Once you enable it, the DBQL views are adjusted automatically so they will reverse the process before the data is viewed, so at the viewing end it should look the same, at least through the standard views.

It's enabled/disabled through DBScontrol flag. Look at Tech Alert 932 for instructions.

Unique query ID is in some 5.0, 5.1, 6.0 and all 6.1 releases.

Some recommendations if you plan to use it:

-- Offload all current DBQL data before making the change

-- Make sure that DBQL-related SQL scripts use standard views

Thanks, -Carrie

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