Archives of the TeradataForum

Message Posted: Tue, 14 Feb 2006 @ 11:42:57 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Hi Raju,

what about creating three join indexes, one for the first select, one for the second and one for the third, and then performing the union? The optimizer should use them instead of base tables and your reports should be more efficient.

Pay attention that if you create a JI on one table, you cannot using MLD or FLD against it in loading new data; the only utility you can use in TPump. Again, ETL processes against a base table on which you have created a JI are a little bit more slow, because Teradata need to refresh the JI too.

Another idea is this: can you convert the union clause to a join? If so, you could create a JI on all your 6 tables, or perhaps only on the bigger tables. The optimizer should use it and then join the JI with the remaining tables.

Hope this helps,

Bye,

Francesca

	<Prev	Next>		<<First	<Prev	Next>	Last>>