Razzak, I've a philosophical question about this methodology.
Using the indicator variables suppresses error 2513 where a null is retrieved. Since one can now turn that error message off, I wonder if there remains any value in using the indicator variables. After all, if one wants to know if a null was retrieved, they can test the target variables instead of the indicator variables. Not using the indicators makes for a simpler syntax. Also, not using the indicators, but turning off the error message, you can still test SQLCODE for -2513 to catch if indeed any nulls were retrieved. I know you have a reason for your syntax choices and it would be educational to know your reasoning in this case. Very Best R:egards, Dennis McGrath -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of A. Razzak Memon Sent: Friday, June 19, 2009 1:46 PM To: RBASE-L Mailing List Subject: [RBASE-L] - Re: Converting from pointer to fetch At 01:45 PM 6/19/2009, Doug Hamilton wrote: >Razzak - You are always very meticulous in predefining and clearing >variables, Doug, I was taught the correct way by the father of R:BASE, who is still the "Main Man" in my life, and most importantly still leads R:DreamTeam, Wayne Erickson. >but I noticed that you didn't predefine the indicator variables for the >FETCH statements. Technically, the INDICATOR variable data type is always an INTEGER, and is handled by the R:BASE engine. A 0 is returned if a value was retrieved. >Apparently they don't need to be predefined, but might that practice come >to bite us in future upgrades? No. You don't need to worry about pre-defining the indicator variables and the data type. >Thanks much, Always a pleasure! Very Best R:egards, Razzak.
