Archives of the TeradataForum
Message Posted: Thu, 31 Jan 2003 @ 00:25:39 GMT
In tuning performance of a query generated by MicroStrategy Desktop, I noticed something that puzzled me regarding the Teradata optimizer: two versions of a query, which appear be functionally equivalent, are having radically different response times. The query is as simple as it gets - an inner join like:
Due to the size of the larger table (70 million rows). In one of many tests, I took the SQL above (generated by MicroStrategy) and re- wrote it manually as:
The query generated by the tool runs in about 1.5 hours, whereas the modified version with the derived table runs in 7 minutes, returning identical results. Regardless of issues like skewing, statistics or other similar considerations:
1. aren't the two queries above equivalent?
2. is there a reason why the optimizer seems to work so differently on each query?
Many thanks in advance.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|