You have to use informat to tell SAS how the date is coming in.  You can use format to change the display to anything you want:   DATA Sample2; informat tdate date7.; format tdate date9.; input cust 1-1 @3 tdate date7. @11 amount; Datalines; 1 01Jan12 20 1 05Jan12 40 1 05Jun12 40 1 16Jul12 60 1 25Nov12 10 2 01Jan12 5 2 10Jan12 4 2 08Jun12 34 2 19Jul12 23 2 23Nov12 45 3 08Jan12 23 3 10Jan12 12 3 08Jun12 78 3 19Jul12 55 3 23Nov12 34 ; ... View more

The code I have works with numeic and character fields.  In the example I use both. ... View more

Here is a solution, the _all_ adds some noise but it'll work.:   data have; infile cards dsd; input Obs     RAWREG$     PVPRISM$     OVRIDE3     DAYS; cards; 1,01,0939,,408 2,,0943,,379 3,09,0029,,170 4,11,0948,, 5,,0948,,59 6,04,6108,,57 7,06,0993,,59 8,06,6111,,59 9,06,0993,,220 10,04,0901,,59 ; proc sql; create table get_max as select *,max(obs) as max_obs from have; proc transpose data=get_max out=tran;by max_obs;var _all_; data want (keep=Variable missing non_missing); set tran(rename=(_name_ = Variable)); Missing=nmiss(of col: ); Non_Missing=max_obs - missing; if variable in ('Obs','max_obs') then delete; run; ... View more

I'm not sure if this qualifies as 'straight forward code' but here is another solutions:   data have; informat date date9.; format date date9.; input MachineNo Date income; cards; 1 01Jan2012 1500 1 02Jan2012 2000 1 27Aug2012 300 2 02Jan2012 1200 2 15Jun2012 50 3 03Mar2012 1000 4 08Apr2012 500 ; proc sql; create table want as select machineno,date,income,daily,weekly,sum(income) as monthly from( select *,sum(income) as weekly from( select *,sum(income) as daily from( select machineno,date,week(date) as week,month(date) as month,income from have) group by date) group by week) group by month; ... View more

Here is another solution:   proc sql; create table want as select distinct symbol, mean(price) as Price_Mean, count(price) as Price_N, min(price) as Price_Min, max(price) as Price_Max from stockprices group by symbol; ... View more

Here is a solution:   data have; input variable $25.; cards; DGM_HIGH_IG_DUR24_10_0 CGM_HIGH_IG_DUR24_12_0 CGM_LOW_IG_DUR24_3_5 ; data want; set have; variable = catx('.',scan(variable,5,'_'),scan(variable,6,'_')) * 34; run; ... View more

You should give more information regarding the calculations.  While I'm sure there are several that know exactly what you want I do no.  ... View more

Roger,   I like that solution.  I doctored mine a little:   data want; set have; by coid; if profit < 0 then count + 1; if first.coid then count = 1; if profit >= 0 then count = 0; do until(last.coid); set prep; by coid; if count = 2 then Lossporin2K2 = 1; if count = 3 then Lossin3Cp2k = 1; output; end; drop count; run;   How you have yours gave me a new perspective on double dow,  I wasn't using it to it's potential.  Thanks! ... View more

I redid lossporin2k2, with the below code you can easily add new variables for 2,3,4,5 consecutive losses:   data have; infile cards dsd; input CoID$ Year Profit; cards; 123,1998,-100 123,1999,-100 123,2000,-150 123,2001,200 456,1999,50 456,2000,10 456,2001,-25 789,1999,-150 789,2000,-100 789,2001,50 789,2002,100 ; data prep; set have; by coid; if profit < 0 then count + 1; if first.coid then count = 1; if profit >= 0 then count = 0; run; data want; do until(last.coid); set prep; by coid; if count = 2 then Lossporin2K2 = 1; if count = 3 then Lossin3Cp2k = 1; end; do until(last.coid); set prep; by coid; output; end; drop count; run; ... View more

Here is another solutions:   data have; input Id dx $ test; datalines; 8210 no 1 8210 yes 20 8210 no 30 8321 yes 11 8321 no 2 ; proc sql; create table want as select /*distinct?*/ a.id, a.dx, b.test from have a full join        have b on a.id = b.id; ... View more

There's several ways to do this, here's another idea:   data good bad; set have; if 2 <= age_pdts <= 13 then output good; else output bad; run; ... View more