![](//teradataforum.com/imgs/ptdfw1.gif) |
![](//teradataforum.com/imgs/ptdfw2.gif) |
Archives of the TeradataForum
Message Posted: Wed, 12 Jun 2002 @ 13:46:11 GMT
Subj: | | Re: Null columns on USI/NUSI |
|
From: | | James A. Judge |
For either, one issue is if you want to use the index column in a join. As we know a NULL value never compares equally (or TRUE) to any
other value; so those rows with NULL values are never included in a join. If you specifying "WHERE column IS NULL" (I am not sure why you
would with a USI for one row, but perhaps with a NUSI column you are checking a date column where NULL means current or most recent) then I
suppose the index could be scanned. But I would assume, if the index subtable was considered by the optimizer, that the normal
considerations for use/non-use of an index would apply; those being the % of rows that have the NUSI NULL value and if using the index
subtable results in having to access less than all the underlying data blocks.
| |