Archives of the TeradataForum
Message Posted: Thu, 24 Nov 2011 @ 17:47:14 GMT
Subj: | | Re: Performance issue after splitting the big table |
|
From: | | Walter, Todd |
A better solution would be to partition the table with PPI on the attribute you used to split it into several parts.
If you look at your explains, you will likely see that it is materializing the UNION prior to performing the rest of the query - effectively
putting the table back together first.
If you have criteria specified against the view which can be used to exclude looking at some of the multiple tables, then if you place CHECK
constraints on those attributes of the tables the optimizer will be able to match check constraints against the conditions in the query and
exclude some of the tables from the UNION - effectively a form of PPI. If more than one is selected, it will still likely have to materialize the
UNION.
|