Archives of the TeradataForum
Message Posted: Wed, 18 Jun 2003 @ 20:21:48 GMT
Subj: | | Re: Consecutive sequence sql logic |
|
From: | | Walter, Todd A |
MDIFF was Teradata syntax prior to ANSI defining the standards in this area. The point of the comment in the manual is to encourage
people to use the ANSI form (despite it being more words). MDIFF is still supported for backward compatability.
The original problem is exactly what MDIFF (and the ANSI version) were designed to do. Those that came up with MDIFF solutions get extra
points!
P.S. I note several examples utilizing INNER JOIN syntax. I wonder out loud if people realize that they are defeating the Teradata
optimizer's join ordering logic when they use INNER JOIN instead of writing the joins in the WHERE clause? In V2R5.1 this will be changed
for cases like those in the notes on this topic (one not mixed with OUTER joins of various forms), but until then INNER join will often get
a less than optimal plan.
|