You could use a passthrough... proc sql; connect using LipSDD00 as cLIP; select      recs into      :sourcecnt from connection to cLIP (      select count_big(*) from Fundamental ); quit; ... View more

What's the definition of &output1?  What do you want to be the indicator for a new line? ... View more

I think your first problem would be to describe the logic you will need.  ... View more

Have you selected the scheduling server? ... View more

You can also use passthrough sql to execute packages on the oracle platform.  If you already have a libname defined (say OraLib) for the oracle connection, you can use the new connect using syntax.  If not, then you'll need to use the expanded connection syntax. proc sql; connect using OraLIb as OraCon; execute (      exec schema.packagename      ) by OraCon; disconnect from OraCon; quit; I haven't tested it...but if your package returns values, then you might could use something like create table work.want as select * from connection to OraCon(      exec schema.packagename ); ... View more

something like this? proc sort data=have; by id class descending date;run; data want; retain QualityPoints num; retain TotalCredits num; if first.id then do;      QualityPoints=0;      TotalCredits=0;      end; if first.class then do;      if Grade='A' then QualityPoints+4;      else if Grade='B' then QualityPoints+3;      else if Grade='C' then QualityPoints+2;      else if Grade='D' then QualityPoints+1;          TotalCredits+Credits;      end; if last.id then GPA=QualityPoints / TotalCredits; run; ... View more

proc sql; create table want as select      t1.patient,      t1.visit_date,      t1.Trt_at_visit,      sum(case when t1.trt_at_visit = t2.trt_at_visit then 1 else 0 end) as Future_Rx_Current_Med,      sum(case when t1.trt_at_visit <> t2.trt_at_visit then 1 else 0 end) as Future_Rx_Different_Med from      have t1      left outer join have t2           on t1.patient = t2.patient           and t2.visit_date > t1.visit_date           and t1.visit_date < t1.visit_date + 180 group by      t1.patient,      t1.visit_date,      t1.trt_at_visit; quit; ... View more

Haven't tested...but here's the normal sql approach. proc sql; create table want as select      t1.customer_id,      t1.purchase_date,      count(distinct t2.product_type) as Products_In_Last_Year from      have t1      left outer join have t2           on t1.customer_id=t2.customer_id           and t2.purchase_date >= t1.purchase_date - 365           and t2.purchase_date < t1.purchase_date group by      t1.customer_id,      t1.purchase_date; quit; ... View more

For the first part, if you're joining by id, then its pretty straightforward depending on which table (if any) would be the "master" table.  If neither, then you might consider: proc sql; create table want as select      coalesce(t1,id, t2.id) as ID,      coalesce(t1.a,0) + coalesce(t2.a,0) as a from      dataset1 t1      full outer join dataset2 t2           on t1.id=t2.id; quit; you will have some issues if id is not unique in either table. For the second part (without testing): proc sql; create table want as select t1.* from (      (select * from sql.a except select * from sql.b)      union      (Select * from sql.b except select * from sql.a) ) t1; quit; ... View more

sorry...no other suggestions.  Perhaps you should refer this to sas support. ... View more

Does your process delete the file once it starts? We had to change to file arrival instead of file exists. ... View more

I haven't used Access in this way in a long while...but I think it had built-in permissions.  You could define permissions to select/update/etc. tables based on user id.  ... View more

You can check any embedded permission...as well as table names to ensure SAS compatibility. ... View more

Ctrl-I will format code for you...at least in a code node in EG. (or Edit>Format Code) ... View more