Stephen, I like your #2 & 2b examples. In a reply I submitted my test for changed columns it had a whole lot of "OR" to test for changes. I will experiment to see if most of the "OR" can be replaced by
one of your constructs. Thanks for the hints. Jim Bentley American Celiac Society [email protected] tel: 1-504-737-3293 >________________________________ > From: Stephen Markson <[email protected]> >To: RBASE-L Mailing List <[email protected]> >Sent: Friday, May 18, 2012 8:49 AM >Subject: [RBASE-L] - Re: Difference between "select into" and "setvar" > >Hi Dennis, > >Yes, you can test if a null was assigned to a variable instead of using an >INDICATOR var. I use both techniques. As always with programming, there's more >than one way to skin a cat. For example, you can use a correlated SELECT >instead of a sub-SELECT in most cases. > >A couple of uses for the IND VAR: > >1. Suppose you have a routine that successively SELECTS INTO different vars >and you want to do the same thing if any is null. If you use the same >indicator variable all the time, code for testing for the NULL can be copied >without regard for the name of the variable SELECTed INTO. > >2. If you are SELECTing into a (large) number of variables and you want to see >if ANY are null you can use: > IF var1 IS NULL OR var2 IS NULL OR var3 IS NULL... THEN >Or, if you have IND VARs vI1, vI2, vI3.... you can use > IF (.vI1+.vI2+.vI3+...)<>0 THEN > >2b. Similarly, if you want to see if ANY of the vars have a non-null value, >you can use: > IF (.vI1*.vI2*.vI3*...)=0 THEN > > > >Regards, > >Stephen Markson >The Pharmacy Examining Board of Canada >416.979.2431 x251 > >From: [email protected] [mailto:[email protected]] On Behalf Of Dennis McGrath >Sent: Friday, May 18, 2012 4:59 AM >To: RBASE-L Mailing List >Subject: [RBASE-L] - Re: Difference between "select into" and "setvar" > >It seems that some users of this list find my statements below to be offensive. >I appoligize for any upset I may have caused. > >Does anyone out there have an example of how they use use the indicator >variables after the data has been retrieved? > >I'm always interested in making my code more efficient, and would welcome any >examples where this is acheived with indicator variables. > >Dennis McGrath >[email protected] >[email protected] > >On Thu, May 17, 2012 at 10:57 AM, Dennis McGrath <[email protected]> wrote: >Now that we can easily turn off the message about a null value being returned, >I find the indicator variables totally unnecessary code bloat. > >In the rbase environment, at least, If I want to know if a null is returned I >can test the actual value returned. In over 25 years of progamming rbase, the >only use I ever had for the indicator variables was to avoid throwing the >error to make tracing easier. > >Dennis McGrath >[email protected] >[email protected] > > > > >
