|
Archives of the TeradataForumMessage Posted: Wed, 23 Apr 2003 @ 16:19:35 GMT
At both my previous and current employer, we have had tremendous success using Join Indexes (both JI and AJI.) We have primarly implemented JIs as a solution to moving OLAP cubes back into the Teradata warehouse. Main things to look out for: 1) Must have a thorough understanding of the tool generating the SQL which will use JIs inorder to limit cases where a JI is not a valid candidate 2) The more columns you incoporate into a JI, the less likely the optimizer is to use the JI because the JI may become 'too fat', and a full table scan of the JI might cost more than just joining the filtered underlying tables 3) If you have a case where you think a JI should be used by the optimizer, and it isn't, then open an incidient via Teradata's Global Incident Management web-site. The implementaiton of JI by Teradata customers is still in the early stages. Costing issues exist. If you'd like, you can contact me off-line, and I will be happy to help. Hope this helps, Clay
| ||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||
Copyright 2016 - All Rights Reserved | ||||||||||||||||||||||||||||||||||||||||||||||||
Last Modified: 15 Jun 2023 | ||||||||||||||||||||||||||||||||||||||||||||||||