 |
 |
Archives of the TeradataForum
Message Posted: Tue, 21 Mar 2006 @ 13:57:41 GMT
| Subj: | | Re: Performance issue on a huge table.. |
| |
| From: | | Victor Sokovin |
Daniel,
| | Even if the optimizer only stores the PI in spool it will still have to read all the blocks from the full table. The Disc I/O saved in
reading a table/index only containing PIs may still be beneficial. | |
Right; but what I do not seem to understand is why a single table JI would make a difference in this case. If we take the query in the original
posting as an example, the optimizer should have the range of PIs straight away, right? The query is on UPI and the WHERE clause only refers to
that UPI. This is the part of the trick I do not quite understand. Does it have to do with the range of PIs (number of rows) the query is supposed
to return? If it's high the optimizer follows a different route where JI happens to be useful?
BTW, I fully agree with what you say. I just hope I have now managed to formulate my question better.
Regards,
Victor
| |
