 |
 |
Archives of the TeradataForum
Message Posted: Tue, 21 Apr 2009 @ 09:30:12 GMT
| Subj: | | Re: 2616: Numeric overflow occurred during computation. |
| |
| From: | | Dieter Noeth |
frank.c.martinez wrote:
| | haven't seen the 13 presentation (I wasn't at Partners last year). Cool, very cool. So is part of the optimization to make the resulting
type in COUNT a DECIMAL(18) or BIGINT? (hah, BIGINT is so funny. I think of a large, economy size palooka from some gangster movie, "BigInt, just
ice the guy.", "Sure, boss, whatever ya say!", "No, BigInt, not with a block of ice!" "Sorry, boss.", "Blam!" "BigInt, he was already dead.",
"Oh. Sorry."). | |
Hi Frank, if you run a session in ANSI mode, then the resulting datatype is a decimal(18) or (38). I like the remark in the manuals:
"ANSI mode uses DECIMAL because tables frequently have a cardinality exceeding the range of INTEGER. Teradata mode uses INTEGER to avoid
regression problems."
So it's "backward compatibility", they'll probably never change it.
Dieter
| |
