However, you could use keepme1 keepme2 ... and then use if nmiss(keepme:)=0;

-Joe

On Thu, Jun 30, 2011 at 11:34 AM, DUELL, BOB (ATTCINW) <bd9439@att.com>wrote:

> You're right of course, this will not work for me (duh). > > Oh well, the brain seems to have started the holiday weekend a bit early. > > Thanks Chang! > > -----Original Message----- > From: Chang Chung [mailto:chang_y_chung@HOTMAIL.COM] > Sent: Thursday, June 30, 2011 9:17 AM > To: DUELL, BOB (ATTCINW) > Subject: Re: OT: Chance to Make SAS-L History: Did You Know That... > > On Thu, 30 Jun 2011 15:43:12 +0000, DUELL, BOB (ATTCINW) <bd9439@ATT.COM> > wrote: > > >Although I have never used this tip, I can think of at least one example > where it might be useful. I have a complex data step that merges several > other data sets on a common BY variable (user_id). I need to guarantee > that the record exists in some data sets, but not all. For example, my > current code looks similar to this: > > > >data permdata; > > merge ds1(in=a) > > ds2(in=b) > > ds3 > > ds4 > > ds5(in=c) > > ds6(in=d); > > by user_id; > > if a and b and c and d; > >run; > > > >Every once in a while I need to add or remove data sets from this > structure. Using this tip, I might be able to simplify this as follows: > > > >data permdata; > > merge ds1(in=keepme) > > ds2(in=keepme) > > ds3 > > ds4 > > ds5(in=keepme) > > ds6(in=keepme); > > by user_id; > > if keepme; > >run; > > you might not! since "if keepme;" above does *not* seem to be equivalent to > "if a and b and c and d" above; it is rather equivalent to "if a or b or c > or d"! > cheers, > chang >