SUM Function Precision in Informix 7.3

Posted in 2000

David Hunt — Tue, 19 Sep 2000 22:52:37 -0400 — source: Informix-list mailing list archive (1991-1998)

Hi, Anyone know how Informix handles the precision of two or more integers when using them with a SUM Function? ie. INTEGER(111111.11) INTEGER(111111.11) SUM = INTEGER(8,2) ? Is there a rule? Where is it documented? TIA, Dave (dhunt@filetek.com)

↪ replying to David Hunt

Art S. Kagel — Wed, 20 Sep 2000 14:44:08 -0400 — source: Informix-list mailing list archive (1991-1998)

David Hunt wrote: > > Hi, > > Anyone know how Informix handles the precision of two or more integers when > using them with a SUM Function? > > ie. INTEGER(111111.11) > INTEGER(111111.11) > > SUM = INTEGER(8,2) ? > > Is there a rule? Where is it documented? Umm, integers have no decimal places! Informix convert all numeric types to DECIMAL for built-in aggregates with precision appropriate to the original operands (ie integer to DECIMAL(16,0), float to DECIMAL(16,2), etc.) then returns the result as that DECIMAL type. If your host variable is of another type (ie integer, short, float, double) a conversion takes place on transmission. Art S. Kagel