|
Archives of the TeradataForumMessage Posted: Sat, 15 Sep 2002 @ 03:34:06 GMT
Hi The plan the optimizer comes up with may depend on the literal value coded in the where clause. e.g. select * from table_x where nusi_col = 'rare_value' ; The above query may use a NUSI on column nusi_column as the selectivity for value 'rare_value' is good. On the other hand, for the query where nusi_col = 'common_value' the explain plan may show a FTS. All this makes sense. Now, if I use a macro or BTEQ with using clause, the query without the literal value is in the request parcel, and the literal value is in the data parcel. The optimizer does *not* see the data parcel. So how did the optimizer do its work without knowing the literal value? Therefore, the plan for the macro may not be the 'most optimal' plan? Kumaran Anantaraman
| ||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||
Copyright 2016 - All Rights Reserved | ||||||||||||||||||||||||||||||||||||||||||||||||
Last Modified: 15 Jun 2023 | ||||||||||||||||||||||||||||||||||||||||||||||||