|
Archives of the TeradataForumMessage Posted: Wed, 29 Jan 2014 @ 20:14:02 GMT
Look at the EXPLAIN for the inner join with ROWID. After Table B is joined to the JI, the spool must be redistributed by the ROWID for table A and joined to that table to pick up the other columns. In the case of Outer Join, the ROWID might be NULL for many rows - in which case processing would be very skewed. The optimizer isn't willing to take that chance.
| ||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||
Copyright 2016 - All Rights Reserved | ||||||||||||||||||||||||||||||||||||||||||||||||
Last Modified: 15 Jun 2023 | ||||||||||||||||||||||||||||||||||||||||||||||||