select specific rows in SAS

Question

I have following data with multiple recording from single patient.

ID                       date           day     result
BELF000000084785    20111110    1   164
BELF000000084785    20111202    2   172
BELF000016833470    20070122    2   226
BELF000016833470    20070522    3   398
BELF000016959442    20080429    1   242
BELF000016959442    20080820    1   126
BELF000016959442    20090225    2   302
BELF000016959442    20090424    2   320
BELF000017061714    20080626    1   152
BELF000017801424    20110803    1   298
BELF000017801424    20110913    2   252
BELF000017801424    20111022    3   234
BELF000024405407    20080218    1   232
BELF000024405407    20080318    2   190
BELF000024405410    20070122    3   380
BELF000024405410    20070218    1   506
BELF000024405410    20070320    2   480
BELF000024405410    20070626    3   346
BELF000024408114    20070423    1   296
BELF000024408114    20070711    3   294

I want to extract the patients with records in day 1 ; if day 1 and day 2 both available and day1 day2 and day3 all available. and through away other. any sas help.

if @vasja's answer is not what you want, maybe you could show what successful output looks like. — Jay Corbett, Sep 25 '12 at 12:12

score 0 · Answer 1 · answered Sep 25 '12 at 10:50

Without duplicities on ID - day level you could just skip latest day:

proc sort data=sample;
by ID day;
run;

data result;
set sample;
by ID;
if not last.ID then output;
run;

What do you want to do with duplicite ID - day record on different date(s)?

score 0 · Answer 2 · answered Sep 25 '12 at 13:49

I think this does precisely what was asked for in the OP. I don't know if that's actually what you want, but that requires further clarification.

data have;
informat date YYMMDD8.;
format ID $17.;
input ID $ date day result ;
datalines;
BELF000000084785    20111110    1   164
BELF000000084785    20111202    2   172
BELF000016833470    20070122    2   226
BELF000016833470    20070522    3   398
BELF000016959442    20080429    1   242
BELF000016959442    20080820    1   126
BELF000016959442    20090225    2   302
BELF000016959442    20090424    2   320
BELF000017061714    20080626    1   152
BELF000017801424    20110803    1   298
BELF000017801424    20110913    2   252
BELF000017801424    20111022    3   234
BELF000024405407    20080218    1   232
BELF000024405407    20080318    2   190
BELF000024405410    20070122    3   380
BELF000024405410    20070218    1   506
BELF000024405410    20070320    2   480
BELF000024405410    20070626    3   346
BELF000024408114    20070423    1   296
BELF000024408114    20070711    3   294
;;;;
run;

data want;
do _n_ = 1 by 1 until (last.ID);
 set have;
 by ID notsorted;
 if first.ID then do; 
    lastday=0; 
    done=0; 
 end;
 if (first.ID) and (day ne 1) then lastday=-99;
 else if day ne lastday+1 then done=1;
 else if not done then lastday=day;
end;
do _n_ = 1 by 1 until (last.ID);
 set have;
 by ID notsorted;
 if first.ID then curday=0;
 curday+1;
 if curday le lastday then output;
end;
run;

Results:

Obs      ID        date day result 
1 BELF000000084785  18941 1 164 
2 BELF000000084785  18963 2 172 
3 BELF000016959442  17651 1 242 
4 BELF000017061714  17709 1 152 
5 BELF000017801424  18842 1 298 
6 BELF000017801424  18883 2 252 
7 BELF000017801424  18922 3 234 
8 BELF000024405407  17580 1 232 
9 BELF000024405407  17609 2 190 
10 BELF000024408114  17279 1 296

select specific rows in SAS

2 Answers2