Have you tried the decode function? This works in Oracle, but have not tested in proc sql, but should be something along these lines: proc sql; create table all_cpt_values as select distinct   a1.id,   a1.claim_id,   a1.icd9,   max(decode(b1.cpt = /*Value for CPT_1*/, 1)) as cpt_1,   max(decode(b1.cpt = /*Value for CPT_2*/, 1)) as cpt_2 from   (   select   a.id,   a.claim_id,   a.icd9   from   table_a a   ) a1,   (   select   b.id,   b.claim_id,   b.cpt   from   table_b b   ) b1 where   a1.id = b1.id   and a1.claim_id = b1.claim_id; group by   a1.id,   a1.claim_id,   a1.icd9 quit; ... View more

Those silent errors are never fun when debugging. Glad you are up and running! -shawn ... View more

Mark, In the ODBC connection setup, what Authentication Method did you select? Mine is set to Kerberos, which will pass the credentials from the PC to the database. I don't have to save the credentials in DM Studio, and when I click Test Connect I get the 'Connection Established' dialog box. -shawn ... View more

You mentioned DM Studio will not save the connection, but I did not see where you right-clicked on Data Source Name and chose 'Save User Credentials' in DM Studio. When they are saved the column labeled 'User Credentials Saved' will change to yes (which one of mine does in the screenshot): ... View more

Does the id that installed DM Studio have write access to the configuration file location? ... View more

You can do this in an expression node. Assuming the account name field is named input_name: /*Pre Expression */ string account_name_1 string account_name_2 hidden integer cnt cnt = 0 /*Expression*/ if instr(`input_name`, ',') > 0 or instr(`input_name`, '&') > 0   begin   if instr(`input_name`, ',') > 0   cnt = parse(`input_name`, ',', account_name_1, account_name_2)   else   cnt = parse(`input_name`, '&', account_name_1, account_name_2)   end else begin   account_name_1 = `input_name`   account_name_2 = null end account_name_1 = trim(account_name_1) account_name_2 = trim(account_name_2) ... View more

When you double click the values in the crosstab, does this take you to section two? Or at least prompt you if you want to go to section two? Here are screenshots which I used to recreate your scenario. This worked as you expected: Creating the section link between the cross tab and section 1 (name of the section in my report): Choose which section to link cross tab to: When I click on the numeric data in the cross tab, I am presented with a prompt to either go to Section 1 or go to an External webpage which was already associated to this particular cross tab. ... View more

Will you provide a sample of the data that is provided to you? I have developed something similar to this using custom categories. Although I want to see how the data is sourced prior to providing a solution. Thanks, -shawn ... View more

This is an another option which will eliminate the Cartesian product and potentially help performance. Here you are joining all of the data with a subset of data that has duplicate telephone numbers. I just took your sample data and it worked great. For the full table, I would put an index on the telephone column since you are joining on it to improve performance. proc sql ; create table Dups_BY_ID as SELECT   a.national_id,   a.telephone FROM   tele_dup a where   a.telephone in (   select   telephone   from   tele_dup   group by   telephone   having   count(*) > 1 )   order by 2 desc; quit; ... View more

Cynthia has valid questions that should be answered to give a more exact answer. Here is pseudo proc sql that should get you going: proc sql;   create table output as   select   em.name,   s.current_salary,   case   when e.profile = 'Pilot' and r.promotion = 0 and r.rating = 1 then s.current_salary * 1.2   when e.profile = 'Pilot' and r.promotion = 0 and r.rating = 2 then s.current_salary * 1.12   when e.profile = 'Pilot' and r.promotion = 0 and r.rating = 3 then s.current_salary * 1.08   when e.profile != 'Pilot' and r.promotion = 0 and r.rating = 1 then s.current_salary * 1.12   when e.profile != 'Pilot' and r.promotion = 0 and r.rating = 2 then s.current_salary * 1.06   when e.profile != 'Pilot' and r.promotion = 0 and r.rating = 3 then s.current_salary * 1.04   when e.profile = 'Pilot' and r.promotion = 1 and r.rating = 1 then s.current_salary * 1.2 + s.current_salary * .1   when e.profile = 'Pilot' and r.promotion = 1 and r.rating = 2 then s.current_salary * 1.12 + s.current_salary * .1   when e.profile = 'Pilot' and r.promotion = 1 and r.rating = 3 then s.current_salary * 1.08 + s.current_salary * .1   when e.profile != 'Pilot' and r.promotion = 1 and r.rating = 1 then s.current_salary * 1.12 + s.current_salary * .1   when e.profile != 'Pilot' and r.promotion = 1 and r.rating = 2 then s.current_salary * 1.06 + s.current_salary * .1   when e.profile != 'Pilot' and r.promotion = 1 and r.rating = 3 then s.current_salary * 1.04 + s.current_salary * .1   end as new_salary   from   rating r,   employeemaster e,   salary s   where   e.employee_id = s.employee_id   and e.employee_id = r.employee_id; quit; ... View more

Ronny, I have attached the job in a Data Management package file as well as the ddf file. Let me know if you have any questions importing it into your environment, or any questions about the job itself. Thanks, -shawn ... View more

Ronny, There is not a straightforward way to do this with just one expression. I have attached a screenshot of a job that will calculate the median: Input: Output: The other variables are ones that were created to assist in the median calculation. -shawn ... View more