devsas, The below code should have fixed the last issue with the day 5, as it is specified for dates between 0 and 7. data test1a;     set desktop.test1;     apt_date = datepart(appointment_date);     format apt_date mmddyy10.;     keep patient_id apt_date; run; proc sort data = test1a;     by patient_id apt_date; run; data test1a;     set test1a;     count + 1;     by patient_id;     if first.patient_id then count = 1; run; data test1b;     set test1a;     where count <=2; /*This is done because there are a couple of people with more than 2 visits and the point of the exercise is only after their first visit, so eliminate all of the others*/ run; data test1c;     set test1b;     by patient_id apt_date;     retain adate;     if first.patient_id then do;     adate = apt_date;     end;     if 0<datdif(adate,apt_date,'act/act')<=7 then do;     flag = 0;     output;     end;     else do;     flag = 1;     output;     end;     drop count adate; run; ... View more

data have2;      set have;      if task = 16 then tempnum16 = "Y";      else tempnum16 = "N";      if task = 17 then tempnum17 = "Y";      else tempnum17 = "N"; run; proc sql; create table want as select distinct clm_id, (min(tempnum16)) as tempnum16, (min(tempnum17)) as tempnum17 from have2 group by clm_id; quit; Normally I would use max rather than min, but Y comes after N, so you use min for the summary of the responses. ... View more

The term inner join implies that you only want to keep all of the records which match in both tables. When I read your query it looked like you wanted a table with all of the facilities and all of the email addresses that you could match, rather than only those which were in both tables. The way that the example code was shown implied that you had an email address for every county, but my guess is that you do not. If this was what you wanted then you should specify left rather than inner join with the facility table being specified first in your statement. ... View more

Insert "label" into the code before replace, this will export the labelled column names from SAS into excel. Mind you if you put labels onto the columns in SAS ie rather than UID you have the label "Unique ID variable", "Unique ID variable" is what will appear in Excel. proc export   data=WORK.xxx   dbms=xls   outfile="xxx.xls"   label   replace; run; If you have variable labels and wanted the SAS variable name to actually be output you could put this before your proc export statement proc datasets lib=work memtype=data;   modify xxx;   attrib _all_ label=' '; run; Hope this helps. ... View more

Or if the goal was to actually only output the line for FSDILX 52 $52,000 then you could specify them in the having statement, otherwise the way the code is currently written I agree with ballardw, proc sql will get mad because you are asking about having something, but not specifying what you want it to have. For reference if you wanted the line you had specified earlier you could use this as the having statement having comm_cd = "FSDILX" and count(ln_no) = 52 and sum(balance) = 52000; ... View more