How to have trigger fire after a (conditional) insert in Oracle? - oracle

Question

So I am trying to have a trigger fire and insert data into various tables. However, the process will differ so i was going to create two separate triggers. I've learned thus far how to have a trigger fire after every insert into a table. How can i have a trigger fire ONLY if the IDs are correlated to a certain primary key (id) from another table? I want the trigger to only fire on survey_cycles correlated to Form_IDs of '777' from the Form table. Form_ID and Survey_Cycle are joined at form_id. The bare bones table structure for both of these tables are below:
*Survey_Cycle:*
survey_Cycle_id
survey_form_id
*Survey_Form:*
Survey_Form_Id
My current trigger code is below:
create or replace TRIGGER Survey_Sample
AFTER INSERT
ON Survey_Cycle
FOR EACH ROW
DECLARE
Survey_Cycle_Id Number;
pSurvey_Cycle_Id Number;
BEGIN
Insert into Survey_Cycle_Sample
(Survey_Cycle_ID, Stat_Method_Id, Create_Dt, Create_User_Id, Modify_Dt, Modify_User_Id, Effective_Dt, Inactive_Dt, Survey_Cycle_Sample_Tx)
Values
(:NEW.Survey_Cycle_Id, 0, trunc(sysdate, 'HH'), 1, null, null, null, null, null);
Insert into Survey_Cycle_Period
(Survey_Cycle_Id, Survey_Cycle_Period_Open_Dt, Survey_Cycle_Period_Close_Dt, Survey_Period_Type_Cd, Create_Dt, Create_User_Id, Modify_Dt, Effective_Dt, Inactive_Dt, Survey_Cycle_Period_Due_Dt, Survey_Cycle_Actual_Close_Dt)
Values
(:NEW.Survey_Cycle_Id, trunc(sysdate, 'HH'), trunc(sysdate + 1), 'Initial', sysdate, 1, null, null, null, sysdate - 1, null);
END;

Answer 1

Perhaps you could create a trigger which calls a stored procedure, and let it decide whether do (or not) additional processing. Here's an example:
Sample tables:
SQL> create table survey_cycle
2 (survey_cycle_id number constraint pk_sc primary key,
3 survey_form_id number
4 );
Table created.
SQL> create table survey_form
2 (survey_form_id number constraint pk_sf primary key);
Table created.
SQL>
A procedure, which - for the ID passed to it - checks whether such an ID exists in the SURVEY_FORM table. If not, SELECT will fail (i.e. raise NO_DATA_FOUND) so the procedure won't do anything. If it exists, it'll execute additional code (such as inserts into two other tables; instead of that, I'm just displaying the appropriate message).
SQL> create or replace procedure p_survey (par_survey_form_id in number)
2 is
3 l_survey_form_id survey_form.survey_form_id%type;
4 begin
5 select survey_form_id
6 into l_survey_form_id
7 from survey_form
8 where survey_form_id = par_survey_form_id;
9
10 dbms_output.put_line('Here goes INSERT INTO survey_cycle_sample and survey_cycle_period');
11 exception
12 when no_data_Found then
13 -- there's no ID like PAR_SURVEY_FORM_ID, so - do nothing
14 null;
15 end;
16 /
Procedure created.
The trigger is very simple:
SQL> create or replace trigger trg_ai_survey_cycle
2 after insert on survey_cycle
3 for each row
4 begin
5 p_survey(:new.survey_form_id);
6 end;
7 /
Trigger created.
SQL>
Finally, testing: 777 is the "existing" ID which will cause additional inserts.
SQL> set serveroutput on
SQL> insert into survey_form values (777);
1 row created.
SQL> --
SQL> insert into survey_cycle values (1, 100);
1 row created.
SQL> insert into survey_cycle values (2, 777);
Here goes INSERT INTO survey_cycle_sample and survey_cycle_period
1 row created.
SQL>