The FULL OUTER JOIN keyword returns all records when there is a match in left (table1) or right (table2) table records.
If there are no matches at all, FULL OUTER JOIN can potentially return very large result-sets because it’s going to include everything from left and right tables!
The OUTER keyword doesn’t make any difference in the query. FULL OUTER JOIN and FULL JOIN are the same.
Basically, it looks like this.
The highlighted part from the image above is our query results when we apply FULL OUTER JOIN between 2 tables.
SELECT a.column2, b.column2 FROM table a FULL OUTER JOIN table b ON a.column = b.column;
I am using alias as well in the syntax. You can use name of the table instead of alias. Alias makes it better to write the query. Learn more about alias in this post:
Column And Table Alias In Postgres
FULL OUTER JOIN Query Example 2 tables
I am using the database for all examples. It is available on my Github public repo
SELECT c.contact_name, o.order_date FROM customers c FULL OUTER JOIN orders o ON c.customer_id = o.customer_id //Output contact_name. order_date .... .... .... "Guillermo Fernández" "1998-05-05" "Jytte Petersen" "1998-05-06" "Michael Holz" "1998-05-06" "Laurence Lebihan" "1998-05-06" "Paula Wilson" "1998-05-06" "Marie Bertrand" null "Diego Roel" null ... ... 832 rows
As we can see there were no matches from the right table, so the query result included the data from left table with null. In this case the FULL OUTER join behaves like LEFT join.
What will happen if we apply LEFT join to 3 tables?
INNER JOIN Query Example 3 tables
SELECT c.contact_name, o.order_date, od.product_id FROM customers c FULL OUTER JOIN orders o ON c.customer_id = o.customer_id FULL OUTER JOIN order_details od ON od.order_id = o.order_id //Output contact_name. order_date. count ... ... ... "Paula Wilson" "1998-05-06" 73 "Paula Wilson" "1998-05-06" 75 "Paula Wilson" "1998-05-06" 77 "Marie Bertrand" null null "Diego Roel" null null ... ... 2157 rows
As we can see we are getting all the data possible from LEFT tables and for right tables, if there are not any match, it simply returns null. If we want to filter this data more, we can add an aggregate function and add GROUP BY clause.
As we saw in the previous query, in this query the FULL OUTER join behaves like LEFT join as well.