Yes, just change the j in the second intnx to be one as this part is linked to start_date not to the do loop: data want (drop=j);   format start_date end_date date9.;   do name="Raj","Ravi","Ram";     do j=0 to 3;       start_date=intnx('day4','06FEB2014'd,j);       end_date=intnx('day4',start_date,1);   /* Think you can change this to day4 if that's whats needed */       output;     end;   end; run; So start date will be 4 day intervals 3 times based on the loop, end_date is 4 days after start_date. ... View more

data want (drop=j);   format start_date end_date date9.;   do name="Raj","Ravi","Ram";     do j=0 to 3;       start_date=intnx('day3','06FEB2014'd,j);       end_date=intnx('day3',start_date,j);   /* Think you can change this to day4 if that's whats needed */       output;     end;   end; run; ... View more

Ah, I see.  Its three day gaps or something similar.  Check out the SAS docs: SAS(R) 9.3 Functions and CALL Routines: Reference You can use DAY3 or whatever your interval is, and change the J loop to be how many times you want it: data want (drop=j);   format start_date end_date date9.;   do name="Raj","Ravi","Ram";     do j=0 to 3;       start_date=intnx('day3','06FEB2014'd,j);       end_date=intnx('day3','06JUN2014'd,j);       output;     end;   end; run; ... View more

Errrm, then just add format into mine? data want (drop=j);   format start_date end_date date9.;   do name="Raj","Ravi","Ram";     do j=0 to 3;       start_date=intnx('week','06FEB2014'd,j);       end_date=intnx('week','06JUN2014'd,j);       output;     end;   end; run; ... View more

Yes, not sure what proc output that's generated from but myself I would do: report to rtf with: ods escapechar='^'; footnote1 j=r "Page ^{thispage}"; Then print the rtf to pdf.  You could however go directly to PDF with the same syntax.  Me, I would also put a breaking section in the report to get a clean break.  So: data mydata;      set mydata;      retain pge;      if _n_=1 then pge=1;      if _n_ = 20 then pge=pge+1; run; proc report data=mydata;      columns pge ...;      define pge / order noprint; ...      break after / pge; ... ... View more

Yes, don't do conditionals within groupings.  Also try to think about process rather than overall e.g: What does my output table need.  Where does the data come from.  What do I need to do on it.  Then break your problem up so you can see the pieces: proc sql;   create table WORK.KPTARREARS_VINTAGEVALUE as   select  *,           case  when MONTH_DATE => 'Jan08' THEN '2008+'                 when =< 'dec04' then 'upto2005'                 when between Jan06 and dec06 then 2006                 else '2007' end as YEAR   from    (             select  distinct                     MONTH_DATE,                     count(*) as COUNT,                     sum(BALANCE_OUTSTANDING) as BALANCE,             from    (                       select  *                       from    GBASEL.BASELJUL14                       where   OPTIMUM_PLATFORM = 'Optimum'                         and   ARREARS_MULTIPLIER >= 3                         and   ARREARS_LIT_STAGE_CODE not in ('L4','L5','L6')                     )             group by MONTH_DATE           ); quit; Now this may not be the best way of coding it, however you can copy out the subqueries into proc sql's of their own and see what data is going in, manipulations etc.  Makes it easier to debug.  Note I am selecting data based on a where in a subquery. This is then grouped to get sums and counts.  This data is then passed into the final select which does the case on the column and decides on the output. /* Note not tested as no test data/required output provided */ ... View more

Sorry, what do you mean?  Page numbers are normally calculated by Word in the footnote by using: footnote1 j=r "Page ^{thispage} of ^{lastpage}"; This is easier than calculating it your self. If however you have to have page numbers in the data then you would need to do: Work out how many pages there will be once rendered - this may not be the same as you think bearing in mind word wrapping/style/borders page size/line size etc.  Then you add that info into your data and output per any other data.  With graphs for instance it may be easy: data _null_;      set graph parameters;      call execute('footnote1 j=r "Page '||strip(put(_n_,best.))||' of ||'strip(put(tot,best.))'||";'); ... run; For tables and listings you could be in for whole mess of coding. ... View more

Hi, Not really getting this as there seems to be quite a few questions and bits in there.  Maybe try breaking it down into individual steps, you can always refactor later into a smaller program.  Here is what I gathered from your requirements to find latest date: data have;   attrib var1 periode format=$200. adddate format=best.;   infile datalines;   input var1 $ periode $ adddate; datalines; ei_isbu_m.csv m 56         namq_nace10_k.csv q 123 ei_cphi_m.csv m 53 ei_isrt_m.csv m 15 namq_aux_lp.csv q 45 ; run; data step1;   set have;   attrib date_publi_1 format=date9.;   if periode="m" then date_publi_1=intnx('MONTH','12JAN2014'd,adddate,'SAME');   else date_publi_1=intnx('QTR','12JAN2014'd,adddate); run; proc sort data=step1;   by descending adddate; run; data want;   set step1 (obs=1); run; ... View more

Hi, Patrick has been nice enough to provide many examples of this.  The !! means concatenate.  What the code is doing is checking the existing length of Pone.  Then if less than 10 create a string of 0's to pad out to 10 length and concatenate these to Phone. Also, it does not matter if you input is character and you want character out, you can still do put(input()) as User24feb has stated: Phone_Text=Put(Input(Phone,Best11.),Z11.); This converts the string to number, uses Z. format to padd it, then puts it back to character and assigns the character result back into your original character format.  So you end up with the original variable with the data padded. ... View more

Hi, /* Note not tested */ proc sort data=have out=a (keep=order period area sales);      by order period area; run; proc sort data=have out=b (keep=order period area comm);      by order period area; run; proc transpose data=a out=t_a;      by order period;      var sales;      id area; run; proc transpose data=b out=t_b;      by order period;      var sales;      id area; run; proc sql;      create table WANT as      select     COALESCE(A.ORDER,B.ORDER) as ORDER,                     COALESCE(A.PERIOD,B.PERIOD) as PERIOD,                     A.NORTH as A_NORTH,                     B.NORTH as B_NORTH,                     A.SOUTH as A_SOUTH,                     B.SOUTH as B_SOUTH,                     A.EAST as A_EAST,                     B.EAST as B_EAST,                     A.WEST as A_WEST,                     B.WEST as B_WEST      from        T_A A      full join     T_B B      on           A.ORDER=B.ORDER      and          A.PERIOD=B.PERIOD; quit; ... View more