Here how you can generate the list of desired table names. /* create sample data */ %macro createSampleData(); %do mm=1 %to 12; %let mm=%sysfunc(putn(&mm.,z2.)); %do yyyy=2009 %to 2018; data loandata_&yyyy.&mm.; set sashelp.class(obs=3); run; %end; %end; %mend; %createSampleData(); /* create macro variable &dslist with list of table names beginning of quarter */ proc sql noprint; select catx('.',libname,memname) into :dslist separated by ' ' from dictionary.tables where libname='WORK' and memname like 'LOANDATA^_%' escape '^' and input(scan(memname,2,'_'),yymmn6.) = intnx('quarter',input(scan(memname,2,'_'),yymmn6.),0,'b') ; quit; %put &=dslist; /* create table want */ data want; length _sourceDS sourceDS $41; set &dslist indsname=_sourceDS; sourceDS=_sourceDS; run; proc print data=want; run; ... View more

Looping involving calendar dates in a macro is much easier if you work with SAS date values (and not human readable values such as 2010 and 10). Therefore, I convert 2010 and 10 to an actual SAS data value using the MDY function. Then you can use the INTNX and INTCK functions as needed to work in terms of quarters (or months or weeks, if desired).   %macro loop(startyear,startmonth); %let date=%sysfunc(mdy(&startmonth,1,&startyear)); %do %while(&date<%sysfunc(today())); /* %put %sysfunc(putn(&date,yymm7.)) %sysfunc(year(&date)) %sysfunc(month(&date));*/ %portfolio_(%sysfunc(year(&date)),%sysfunc(month(&date))) %let date=%sysfunc(intnx(quarter,&date,1,b)); %end; %mend; %loop(2010,10) You would probably also be wise to re-configure this code so that you are always working with actual valid SAS date values, rather than discrete year and discrete month used as &date_input.&month. A SAS date value can be formatted and used in file names as 201907, rather than having to append discrete year and discrete month onto the end a file name. ... View more

Please provide some example input data and what the result for that input data should look like. You might also define exactly what the denominator and numerator should be when creating the proportions or percentages.   And do you actually need a dataset, for further manipulation, or a report for people to read?   Consider this as one example of a report that sums a variable by region  and determines the percentage of that sum as the percent of the whole. You should have the sashelp.shoes data set available. proc tabulate data=sashelp.shoes; class region ; var sales; table region all='All Regions', sales* (sum pctsum); run; This procedure actually will create output data sets as well that can be processed though some folks don't like them.   ... View more