try this one to get the first table: proc sql; create table work.want as select      t1.year,      t1.age,      t1.x1,      t1.x2,      sum(t2.x1) as t2x1,     sum(t2.x2) as t2x2,     sum(t2.x1)/sum(t2.x2) as x3 from      have t1      left join have t2           on t1.age=t2.age                and t2.year between t1.year and (t1.year+4) group by      t1.year,      t1.age,      t1.x1,      t1.x2; quit; ... View more

I think that approach doesn't meet the stated objectives of 5 year increments. proc sql; create table work.want as select      t1.year,      t1.age,      t1.x1,      t1.x2,      sum(t2.x1)/sum(t2.x2) as x3 from      have t1      left join have t2           on t1.age=t2.age                and t1.year ge t2.year                and t1.year+5 lt t2.year group by      t1.year,      t1.age,      t1.x1,      t1.x2; create table want2 as select      t1.*,      1/20*ln(t2.x3/t1.x3) as x4 from      want t1      left join want t2           on t1.age=t2.age                and t1.year-20=t2.year; quit;    ... View more

proc sql; delete from work.have where group in (select group from work.have where diff=0); quit; ... View more

I think the general solution is to build a dataset consisting of the days you consider to be "work" days and then summing those.  There's a support doc on how to do it. 26044 - Determine the number of business days between two SAS dates proc sql;   create table final_sql as   select startdt format=date9.,   stopdt format=date9.,   (select count(*)   from calendar   where dt between stopdt and startdt   and type = 'Workday') as workdays   from test; quit; ... View more

proc sql; delete from lib_ref1.ods_customer t1 where exists (select * from lib_ref2.cust where id1=t1.id1 and id2=t2.id2); quit; ... View more

If you're using Enterprise Guide, there is an option that tells EG to "Automatically open data or results when generated".  Go to Tools>Options>Results General. ... View more

You have to manually match up the columns in the two clauses.  You shouldn't include the identity column but only those that you have in your sas table.  I may have the table names reversed...but it sounds like you're inserting into the sql table and selecting from the sas dataset. proc sql; insert into sqllibname.sqltable (      columns that match the sas dataset )      select           column names in the same order as above      from           saslibrary.sastable; quit; ... View more

Since you have an identity column in sql you need to specify the columns that you're inserting into: proc sql; insert into have (      col1,      col2,      ....      ) select      col1,      col2,      ... from want; quit; ... View more

I don't understand why you have a Y for ID 2.  Here's my best guess for what you're trying to do: proc sql; create table want as select        t1.id,     t1.date,     t1.spend,     sum(t2.spend) as Rolling_Spend,     case when sum(t2.spend) >=50 then 'Y' end as Flag from        have t1     left  join have t2         on t1.id=t2.id             and intck('month',t2.date,t1.date) between 0 and 2 group by t1.id, t1.date, t1.spend order by t1.id, t1.date; quit; id date spend Rolling_Spend Flag 1 01Jan2010 20 20 1 01Feb2010 10 30 1 01Mar2010 0 30 1 01Apr2010 10 20 1 01May2010 40 50 Y 2 01Mar2009 20 20 2 01Apr2009 10 30 2 01May2009 10 40 2 01Jun2009 10 30 2 01Jul2009 5 25 2 01Aug2009 15 30 ... View more