Hi, Could you please clarify what input data do you have? Is it like a matrix? (If so, is the 4th observation in 1st row missing so that it would be (3X4) matrix. Or do you just want to make data set like below from no initial data? ... View more

Hi, At least it seams to me that the problem is that you have 2 names for libname 'sc' and 'period1'. Maybe you should try simply libname period1 'c:\Cwa'; ... View more

Maybe option firstobs=2 is what you need ... View more

You can try this, maybe that is what you need: data test; input Date date9. Firm $ Var Avg_necessary $9.; format date date9.; datalines ; 1-Jan-08 ABC 1 1 1-Feb-08 ABC 2 2 1-Mar-08 ABC 3 3 1-Apr-08 ABC 4 4 1-May-08 ABC 5 5 1-Jun-08 ABC 6 6 1-Jul-08 ABC 7 7 1-Aug-08 ABC 8 8 1-Sep-08 ABC 9 9 1-Oct-08 ABC 10 10 1-Nov-08 ABC 11 11 1-Dec-08 ABC 12 12 1-Jan-09 ABC 13 (1+13)/2 1-Feb-09 ABC 14 (2+14)/2 1-Mar-09 ABC 15 (3+15)/2 1-May-09 ABC 16 (4+16)/2 1-Jun-09 ABC 17 (5+17)/2 1-Jul-09 ABC 18 (6+18)/2 1-Jan-08 XYZ 21 21 1-Feb-08 XYZ 22 22 1-Mar-08 XYZ 23 23 1-Apr-08 XYZ 24 24 1-May-08 XYZ 25 25 1-Jun-08 XYZ 26 26 1-Jul-08 XYZ 27 27 1-Aug-08 XYZ 28 28 1-Sep-08 XYZ 29 29 1-Oct-08 XYZ 30 30 1-Nov-08 XYZ 31 31 1-Dec-08 XYZ 32 32 1-Jan-09 XYZ 33 (21+33)/2 1-Feb-09 XYZ 34 (22+34)/2 1-Mar-09 XYZ 35 (23+35)/2 1-May-09 XYZ 36 (24+36)/2 1-Jun-09 XYZ 37 (25+37)/2 1-Jul-09 XYZ 38 (26+38)/2 ; run; proc sort data=test; by firm date; run; data test2; set test; by firm; if first.firm then counter=0; counter+1; lag12=lag12(var); if counter le 12 then avg=var; else avg=(var+lag12)/2; run; You can drop counter and lag12 from final data of course. ... View more

Not sure if I understand what do you want to calculate. Would it be just one number - value for Jan09 =(1+13)/2 or you want averages for each following month? ... View more

Another simple way how to do that: data two; set one; where x=int(x); run; ... View more

Hi, I think adding these couple of lines works as you wished: DATA SET2 (drop=num); SET SET1; BY OBS; RETAIN STR1; LENGTH STR1 $ 100.; LENGTH STR2 $ 100.; if first.obs then num=0; num+1; IF (FIRST.OBS AND LAST.OBS) THEN DO; STR1 = NAME; END; ELSE DO; IF FIRST.OBS THEN DO; STR1 = NAME; END; ELSE IF num le 10 then DO; STR2 = STR1; STR1 = ' '; STR1 = TRIM(STR2)||','||TRIM(NAME); END; END; IF LAST.OBS THEN DO; OUTPUT; STR1 = ' '; END; DROP NAME STR2; run; ... View more

Could you please show with some example lines how your data look like and what you want to get? It sounds like you have to use Proc Format. ... View more

Try using "." after calling macro: CREATE TABLE &A._MASTER_TO_DUMP AS SELECT DISTINCT (SUM(ACTUAL_DISBURSAL/10000000)) AS ... View more

Simple code that works: data have; input Var1 $; datalines; qwer asdf zxcv ; run; proc transpose data=have out=want (drop=_name_); var var1; run; ... View more

Have never used this procedure, but if no other ideas, check: http://support.sas.com/documentation/cdl/en/etsug/60372/HTML/default/etsug_datasrc_sect006.htm So it could be %process (week); %process (day); But it is only a guess, so hopefully somebody with more experience will answer. Good luck with your work! ... View more

Thank you, this works as expected! ... View more

Changed the code a bit: you can add one extra variable that counts how many times the same situation is repeated. So even if it is repeated also in b, this should delete only those observations that match and not all of them: proc sort data=a ; by cust prod cost; run; proc sort data=b; by cust prod cost; run; data a1; set a; by cust prod cost; if first.cost then count=0; count+1; run; data b1; set b; by cust prod cost; if first.cost then count=0; count+1; run; data c (drop = count); merge a1 b1 (in=b); by cust prod cost count; if b then delete; run; ... View more

Hi, This code could help, but only if in the dataset B there are no repeated observations. proc sort data=a ; by cust prod cost; run; proc sort data=b; by cust prod cost; run; data c; merge a b (in=b); by cust prod cost; if first.prod and first.cost and b then delete; run; ... View more