Materialized view to ensure data integrity over multiple tables

Question

Hello,
I have a problem and I am not able to solve it. Partially, It is because of my poor SQL skills. I have three tables and I am using one sequence to enter data into them.
What I am tying to do is to create a materialized view ( complete or fast, whichever) with refresh on commit option to check that each table contains unique data in comparison to other.
I am posting code so you can get along:
CREATE TABLE table_1 (
ID NUMBER PRIMARY KEY
CREATE TABLE table_2 (
ID NUMBER PRIMARY KEY
CREATE TABLE table_3 (
ID NUMBER PRIMARY KEY
INSERT INTO table_1 VALUES (1);
INSERT INTO table_1 VALUES (2);
INSERT INTO table_2 VALUES (3);
INSERT INTO table_2 VALUES (4);
INSERT INTO table_3 VALUES (5);
INSERT INTO table_3 VALUES (6); I want to write create a materialized view that will give me output only in case that there are same values in two different tables. I got this far.
CREATE MATERIALIZED view mv_test
REFRESH COMPLETE ON COMMIT
AS
SELECT count(1) ROW_COUNT
FROM dual
WHERE EXISTS (
     SELECT a.id
     FROM table_1 a
     WHERE a.id IN(
         SELECT b.id
         FROM table_2 b))
OR EXISTS (
    SELECT a.id
     FROM table_1 a
     WHERE a.id IN
             (SELECT c.id
             FROM table_3 c))
OR EXISTS (
    SELECT b.id
         FROM table_2 b
    WHERE b.id IN
             (SELECT c.id
             FROM table_3 c));
    ALTER MATERIALIZED VIEW mv_test
    ADD CONSTRAINT cs_mv_test
    CHECK (row_count = 0) DEFERRABLE; This sql statement itself returns no rows if my logic is correct. And in case there were some duplicate rows in two different table, it would return 1 and constraint would throw an error.
However, I cannot create this with ON COMMIT option. When I try to compile I get:
ORA-12054: cannot set the ON COMMIT refresh attribute for the materialized view I went through documentation, tried creating mat_view logs etc.
I know that one of the mistakes is that I am referencing dual table and I am not sure if I can use EXISTS.
Unfortunately, my SQL wisdom ends here. I need help rewriting the sql, so it would work in materialized view with refresh on commit option. Please, help!!!
I know that since I am using a sequence there is little chance that same value will get into two different tables, but I would like to perform somekind of check.
Thank you in advance.

Answer 1

>
I know that since I am using a sequence there is little chance that same value will get into two different tables, but I would like to perform somekind of check.If you are certain that you control all the inputs to the table and you are definitely using one sequence to insert into all three tables then there is physically no possible way you will get duplicate values across tables.
Writing something to check if this is the case would almost be like writing something to verify that 1+1 really does equal 2 in 100% of cases.
if you must, however. consider something similar to the following which may be more performant:
select *
  from table_1 t1
  full outer join table_2 t2 on (t1.id = t2.id)
  full outer join table_3 t3 on (t1.id = t2.id
                                 or
                                 t2.id = t3.id)
where t1.id+t2.id+t3.id not in (t1.id,t2.id,t3.id);