I'm sorry to hear about your pointaphobia. Is it covered by your medical insurance?

The advantage of POINT in this example is that only the selected records are read twice. Which I suppose could provide a small performance advantage.

On 2/23/12, Keintz, H. Mark <mkeintz@wharton.upenn.edu> wrote: > Here's one that does the same, but indulges my compulsion to avoid the > POINT= option. > > If you don't want to output any records for cases with only 1 record in the > group (the singe age=16 in sashelp.class), then drop the "BY AGE" and "IF > FIRST.AGE=1..." statements. But keep the "SET AGE" statement. > > Regards, > Mark Keintz > > > data want ; > set have; > > by age; > ** Output singletons **; > if first.age=1 and last.age=1 then output; > > set have (firstobs=2 keep=age rename=(age=nextage)); > by nextage; > ** Now output next-to-last records **; > if last.nextage and age=nextage then output; > run; > > > > Mark Keintz > Wharton Research Data Services > 215/898-2160 > > -----Original Message----- > From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Data > _null_; > Sent: Thursday, February 23, 2012 2:41 PM > To: SAS-L@LISTSERV.UGA.EDU > Subject: Re: outputting the next to last observation per group > > Here AGE in SASHELP.CLSSS equal PERSON. > > When last is also first what do you want to do? I used last. > > proc sort data=sashelp.class out=class; > by age; > run; > proc print; > run; > data next2Last; > _n_ = 0; > do until(eof); > do _n_ = _n_+1 by 1 until(last.age); > set class(keep=age) end=eof; > by age; > end; > point = _n_-(last.age*not first.age); > obs = point; > set class point=point; > output; > call missing(of _all_); > end; > stop; > run; > proc print; > run; > > On 2/23/12, Alison <alisontetler@yahoo.com> wrote: >> Hi - I think this should be an easy task but can't quite get it. I >> have multiple observations per person (series of births) and want to >> select the next to last observation in the dataset per person. How do I >> do this?? >> Thanks! >> Alison >> >