 |
 |
Archives of the TeradataForum
Message Posted: Tue, 11 Aug 2004 @ 00:08:07 GMT
| Subj: | | Re: How to read Explain Outputs? |
| |
| From: | | Christie, Jon |
If the optimizer chooses to user a join or hash index, you'll see the join or hash indexes name in the explain. If the optimizer chooses to
use a secondary index it will say something about it in the explain and it will identify the index by a number evenly divisible by 4. (4 is the
1st secondary index, 8 is the second, and so on.) If the optimizer chooses to use the primary index it will say so. (And it will usually happen
on a single-amp step.
You can't guarantee the use of a secondary index. But as you gain experience, you'll probably get a feel for the kinds of situations where the
optimizer is likely to like your index. . . . I know, that's evasive, but the whole story is very long...
When mload is in its application phase (this is where the target tables are worked on) only sessions using an access lock can get at the target
tables. If you're one of those guys, you might encounter a problem due to a duplicate row, but it's not likely.
Jon
| |
