|
Archives of the TeradataForumMessage Posted: Fri, 25 Apr 2003 @ 16:33:34 GMT
I wanted to add on to Dennis's point below, about the optimizer using base table statistics for single table join indexes. There is one exception to this is V2R5, which is not currently documented. In the case of a sparse index (which is implemented as a join index), you must collect statistics on the sparse join index column that defines the "sparseness", or the optimizer will not chose to use the sparse index. For example, if your sparse index looks like this: create join index cabsparse as select c_name,c_custkey,c_acctbal,c_nationkey from customer where c_nationkey = 24 primary index(c_name); You would need to collect statistics on the join index column c_nationkey. Also, in answer to the previous question from Yvonne...Can I use RANK and CASE in a join index? Unfortunately, no. Current join index functionality does not support either in the create join index statement. Thanks, --Carrie
| ||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||
Copyright 2016 - All Rights Reserved | ||||||||||||||||||||||||||||||||||||||||||||||||
Last Modified: 15 Jun 2023 | ||||||||||||||||||||||||||||||||||||||||||||||||