Archives of the TeradataForum

Message Posted: Mon, 10 May 2004 @ 18:48:38 GMT

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

<-- Anonymously Posted: Monday, May 10, 2004 14:39 -->

First you should analyze the existing query.

Check - which step is taking more time?

Check - would collecting statistics on the column help you?

Check - would another USI/SI/NUSI help you?

Check - the existing data distribution of the step.

Check - if any redistribution is occurring.

Check - if a dummy join can help in resolving the problem causing join.

Join Index is a considered as a last solution when most methods fail of a join query.

I don't say it is not a good but you might have to consider a lot n lot of things to have JI in your warehouse. Fr more info on JI, the best place is the Teradata documentation (Database Design volume).

I don't mind saying that if you can figure out the problem of the existing query, then you can avoid writing a join index as well.

Thanks.

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