Re: AW: Performance Impact of SQL Views (was: SQL Statement length)

The performance issue came up because the original view was being
re-rendered
too many times per day (in the tens of thousands). Changing it to an MQT
cut this down to being rebuilt at most once per day. A definite
improvement.

I can definitely see how running simple queries against a single table
(MQT) would perform better than having thousands of JOBS dynamically
"joining" 14 tables at runtime by "opening" a VIEW.

When you began this thread, I thought you were questioning the performance
of SQL VIEWS vs. embedding essentially identical SQL in HLL programs or SQL
stored procedures. That would be an apples to apples comparison. And I
suspect those comparisons would yield little if any difference in
performance.

Creating an MQT would save a lot of query processing associated with
dynamic joins at runtime. Seems like a good design.