Avoid Joining Table Multiple Times

Question

I have two tables EMPLOYEE and EMPLOYEE_HOURS.
EMPLOYEE contains EMPLOYEE_ID
EMPLOYEE_HOURS contains EMPLOYEE_ID, HOURS, TYPE
So the EMPLOYEE_HOURS table will contain multiple entries for a given period with hours for types 1, 2 and 3.
I need to display all this on a single line. My query today is
select e.employee_id, eh1.hours, eh2.hours, eh3,hours
from employee e
inner join employee_hours eh1 on e.employee_id = eh1.employee_id and eh1.type = 1
inner join employee_hours eh1 on e.employee_id = eh1.employee_id and eh1.type = 2
inner join employee_hours eh1 on e.employee_id = eh1.employee_id and eh1.type = 3This only works if I know the number of types and their codes. I know I could create a procedure and build the query dynamically, but I’m wondering if there’s some other SQL and/or PL/SQL that can handle this in a single query?
Thanks,
Mark

Answer 1

If you don't know the number of codes in advance, then you don't know the number of columns in the result set in advance. Since a single SQL statement cannot return different numbers of columns on different executions, that almost always means that you need to use dynamic SQL to generate the SQL statement at runtime. It is possible, as well, to write a generic pipelined table function that returns a different number of rows but that is generally vastly more complicated (there have been examples posted here in the forums if you really want to go down that path).
If you know the number of codes (or you can return a fixed number of columns that is known in advance), you don't need to know the individual code values
SELECT employee_id,
       MAX( CASE WHEN rnk = 1 THEN hours ELSE null END) hours_1,
       MAX( CASE WHEN rnk = 2 THEN hours ELSE null END) hours_2,
       MAX( CASE WHEN rnk = 3 THEN hours ELSE null END) hours_3
  FROM (SELECT employee_id, hours, rank() over (partition by employee_id order by type_id) rnk
         FROM employee)
GROUP BY employee_idJustin