Date:         Wed, 12 Jan 2000 03:01:11 +0100
Reply-To:     Fabrizio Carinci <carinci@CMNS.MNEGRI.IT>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
Comments:     To: "Lucas, Peter (Syd)" <Peter.Lucas@corpmail.telstra.com.au>
Comments:     cc: SASL <SAS-L@VM230.akh-wien.ac.at>
From:         Fabrizio Carinci <carinci@CMNS.MNEGRI.IT>
Subject:      Re: Problems with data step.
Comments: To: SAS-L@LISTSERV.VT.EDU
In-Reply-To:  <034452FA79E2D211A2DC0008C724EDA501FA3D8F@ntmsg0024.corpmail.telstra.com.au>
Content-Type: TEXT/PLAIN; charset=US-ASCII

Pete,
that means you will calculate profeta only when you meet the first record
of a new combination of those variables.
The by statement following a SET does create first.anything for you.
(quite a singular sintax for SAS, but there are many singularities in this
language).

However,
in the case of Angelo Menna, I apologise for a likely misunderstanding:
he was perhaps referring to a single combination per subject (i.e.
distinct values), which in case may be solved by something like:

data pippo;
input PATIENTC TRT1 TRT2 TRT3 TRT4;
cards;
131130         1         2         5         90
131698         90    . . .
131718         2         3         90       90
131762         2         4         90       90
131768         2         4         90       90
131771         90     . . .
131804         90       90  . .
131806         2         3         4  .
131923         1         90       90  .
131975         2         90       90  .
;
run;

data galeone (keep=profeta);
set pippo;
length profeta $200;
array trt[4] trt1-trt4;

do i=1 to 4;
 repeat=0;
 do j=(i+1) to 4;
  if trt[i]=trt[j] then repeat=1;
 end;
 if i=1 and repeat=0 then profeta=trim(left(trt[i]));
 else do;
  if repeat=0 and trt[i]^=. then do;
   if trim(left(profeta))^='' then do;
    profeta=trim(left(profeta))||","||trim(left(trt[i]));
   end;
   else profeta=trim(left(trt[i]));
  end;
 end;
end;
run;
proc print;
run;


Again, this might be easily generalized to n variables...
Ciao
Fabrizio

On Wed, 12 Jan 2000, Lucas, Peter (Syd) wrote:

> Fabrizio,
>
> I'm new to SAS and am curious about the follow code
>
> if first.trt1 & first.trt2 & first.trt3 & first.trt4;
>
> Does this mean we will only calculate profeta if its the first record in
> each column ?
>
> Pete
>
> > -----Original Message-----
> > From:       Fabrizio Carinci [SMTP:carinci@CMNS.MNEGRI.IT]
> > Sent:       Wednesday, January 12, 2000 11:47 AM
> > To: SAS-L@AKH-WIEN.AC.AT
> > Subject:    Re: Problems with data step.
> >
> > Dear Angelo,
> > please try to use the following:
> >
> > =======
> >
> > proc sort data=pippo;
> > by trt1 trt2 trt3 trt4;
> > run;
> >
> > data galeone (keep=profeta);
> > set pippo;
> > by trt1 trt2 trt3 trt4;
> > if first.trt1 & first.trt2 & first.trt3 & first.trt4;
> > profeta=trim(left(trt1))||","||trim(left(trt2))||","||trim(left(trt3))||",
> > "||trim(left(trt4));
> > run;
> >
> > proc print;run;
> >
> > ======
> >
> > You may easily generalize from that by using do loops, or use proc means
> > instead of proc freq, with the 'type' selection on.
> >
> > Hope this helps,
> > Ciao
> > Fabrizio
> >
> > On Wed, 12 Jan 2000, Angelo Menna wrote:
> >
> > > Dear All,
> > > please consider the following data:
> > >
> > > PATIENTC TRT1 TRT2 TRT3 TRT4
> > > 131130         1         2         5         90
> > > 131698         90
> > > 131718         2         3         90       90
> > > 131762         2         4         90       90
> > > 131768         2         4         90       90
> > > 131771         90
> > > 131804         90       90
> > > 131806         2         3         4
> > > 131923         1         90       90
> > > 131975         2         90       90
> > >
> > > I'd like to have a new var (TERAPIA) to obtain distinct values of
> > trt1-trt4
> > > separated by comma
> > > (e.g. for PATIENTC=131718 TERAPIA='2,3,90')
> > >
> > > I tried this code (but it does not work):
> > >
> > > data pluto ;
> > > length terapia temp1 $30 temp2 $2 ;
> > > set  pippo ;
> > >
> > > array baudo trt1-trt4 ;
> > >
> > > do i=1 to 4;
> > >  if i=1 then
> > >   do;
> > >    terapia=trt1 ;
> > >    temp2=trt1 ;
> > >   end;
> > >  else
> > >   do;
> > >    if baudo(i) then
> > >     do;
> > >      if temp2^=baudo(i) then
> > >       do;
> > >       temp1=terapia||','||baudo(i) ;
> > >       temp2=baudo(i) ;
> > >       terapia=temp1;
> > >       end;
> > >     end;
> > >   end;
> > > end;
> > > run;
> > >
> > >
> > > Any help is appreciated.
> > > TIA, Angelo.
> > >
> >
> > ==========================================================================
> > =====
> > Fabrizio Carinci, Statistician
> > Head, Unit of Statistics and Information Systems,
> > Department of Clinical Pharmacology and Epidemiology
> > Consorzio Mario Negri Sud, Strada Nazionale-66030 S.Maria Imbaro
> > (Chieti)-ITALY
> > Web: http://www.cmns.mnegri.it
> > --------------------------------------------------------------------------
> > -----
> > TEL: 39-0872-570265   FAX:  39-0872-578240
> > E-MAIL:carinci@cmns.mnegri.it
> > ==========================================================================
> > =====
>

 ==============================================================================
Fabrizio Carinci, Statistician
Head, Unit of Statistics and Information Systems,
Department of Clinical Pharmacology and Epidemiology
Consorzio Mario Negri Sud, Strada Nazionale-66030 S.Maria Imbaro (Chieti)-ITALY
Web: http://www.cmns.mnegri.it
-------------------------------------------------------------------------------
TEL: 39-0872-570265   FAX:  39-0872-578240     E-MAIL:carinci@cmns.mnegri.it
 ==============================================================================