Question

I have posted this question before. But now my project manager came back and gave me a new set of instructions. Again, I am a bit lost now and tried to fix it to some extend.

I am working on a trigger which needs INSERT INTO with WHERE logic.

I have three tables.

Absence_table:

-----------------------------
|  user_id | absence_reason |
-----------------------------
|  1234567 |   40           |
|  1234567 |   50           |
|  1213    |   40           |
|  1314    |   20           |
|  1111    |   20           |
-----------------------------

company_table:

-----------------------------
| user_id  | company_id     |
-----------------------------
| 1234567  |  10201         |
| 1213     |  10200         |
| 1314     |  10202         |
| 1111     |  10200         |
-----------------------------

employment_table:

-------------------------------------------
| user_id  |   emp_type    |  employee_id |
-------------------------------------------
| 1234567  |   Int         |    1         |
| 1213     |   Int         |    2         |   
| 1314     |   Int         |    3         |
| 1111     |   Ext         |    4         |
-------------------------------------------

and finally I have the table out where data should be going only who have emp_type = Int in employment_table and have company_id = 10200

out:

-------------------------------------------
| employee_id | absence_reason |  user_id |
-------------------------------------------
|  1          |    40          |  1234567 |
|  1          |    50          |  1234567 |
|  2          |    40          |  1213    |
|  3          |    20          |  1314    |
-------------------------------------------

Here is my trigger:

CREATE OR REPLACE TRIGGER "INOUT"."ABSENCE_TRIGGER" 
   AFTER INSERT ON arc_hrcs.absences_data 
   FOR EACH ROW
DECLARE
BEGIN
  CASE
      WHEN UPDATING THEN
           MERGE INTO out o USING DUAL ON (out.user_id =:NEW.user_id)
           WHEN MATCHED THEN UPDATE SET
                      out.employee_id = (SELECT employee_id FROM employment_table WHERE user_id = :NEW.user_id),   
                      out.absence_reason = :NEW.absence_reason,
                      out.user_id = :NEW.user_id
           WHEN NOT MATCHED THEN 


     insert into out (absence_reason, employee_id)
select :NEW.absence_reason, e.employee_id
  from employment_table e 
       inner join company_table c
               on c.user_id = e.user_id
 where e.user_id = :NEW.user_id
   and e.emp_type = 'INT'
   and c.company_id = '10200';
  END CASE;
END absence_trigger;

I can’t figure out how to change the code after WHEN NOT MATCHED THEN code according to the merging syntax. A bit of guide would help me out 🙂

Thanks in advance.

Answer 1

My previous answer was a bit too quickly.

I would rewrite the merge as:

MERGE INTO out o USING (select e.employee_id, :NEW.user_id as user_id
                        from employment_table e 
                        inner join company_table c
                           on c.user_id = e.user_id
                        where e.user_id = :NEW.user_id
                             and e.emp_type = 'INT'
                             and c.company_id = '10200') S
ON (out.user_id =S.user_id)
WHEN MATCHED THEN UPDATE SET
     out.employee_id = s.employee_id,   
     out.absence_reason = :NEW.absence_reason,
     out.user_id = :NEW.user_id
WHEN NOT MATCHED THEN 
     insert (absence_reason, employee_id)
     values (:NEW.absence_reason, S.employee_id);