Dennis, I have also been considering doing away with the Indicator Variables. I routinely code with indicator variables just to avoid the error messages; however, if I need to verify if a variable is null, I normally test the variable itself, rather than the indicator variable. Like you, I am wondering if there is a downside to doing away with indicator variables. Maybe we can request an option to turn off Indicator Variable error messages at the database settings level, so we don't have to constantly do it at the program level. Javier, Javier Valencia, PE 913-829-0888 Office 913-915-3137 Cell 913-649-2904 Fax <mailto:[email protected]> [email protected] _____
From: [email protected] [mailto:[email protected]] On Behalf Of Dennis McGrath Sent: Wednesday, June 08, 2011 1:29 PM To: RBASE-L Mailing List Subject: [RBASE-L] - Re: INDICATOR var with aggregate functions I personally am moving away from using indicator variables at all. I turn off the error message instead If I really need to know if a variable is null, I test it. Testing the indicator variable is an unnecessary step. I would dearly like to know when this could trip me up. I haven't tested an indicator variable in all my 25+ years with RBASE Dennis McGrath Software Developer QMI Security Solutions 1661 Glenlake Ave Itasca IL 60143 630-980-8461 [email protected] _____ From: [email protected] [mailto:[email protected]] On Behalf Of [email protected] Sent: Wednesday, June 08, 2011 1:01 PM To: RBASE-L Mailing List Subject: [RBASE-L] - Re: INDICATOR var with aggregate functions Mike: I'm with Albert. I never use indicator variables in aggregate math calcs because it will always return something, at least a 0. It will never be null unless you did something like "SELECT MAX(textcolumn)" and the entire column was null. Karen In a message dated 6/8/2011 11:17:16 AM Central Daylight Time, [email protected] writes: Do I understand correctly that it is not necessary/advisable to use an INDICATOR variable with aggregate functions such as SUM, COUNT, etc.? The reason I'm asking is that in testing v9.1 - 64 some of my programs that use this scenario stops R:BASE in its tracks. Here's an example where R:BASE gets to this command and then just drops back to the R>. SELECT COUNT(*) INTO VCNT INDICATOR VIND FROM REJ_VIEW + WHERE DISPDATE = .VPREVIOUS AND IQL_SMRY = .VM AND IQL_FLAG = .VQ + AND SEQNUM NOT IN (SELECT RJCT_SEQNUM FROM REWIND_EXCL) + AND REJ_AREA = .VINT AND HOLDSTAT = .VI Taking the INDICATOR out makes R:BASE happy again.
