Question

I am trying to write a SQL statement between two tables where there is a has-many relation between two tables going one way and returns only the rows with more than one relation. I’m probably phrasing this horribly so here’s the example.

Let’s say table A has a column “key_a” that is a primary key for Table A. Table B has two columns. The first is called “key_b” which is a primary key for Table B and a second column called “key_a” which is a foreign key with a relation to “A.has_a”. Application logic has been implemented to ensure that While a row entry in Table A can have a relation from 0 to n rows in Table B, rows in Table B have a relation to exactly one row in Table A.

How can I write an SQL query that will return row entries in “A.key_a” that have more than one row in Table B mapping a relation to it through the “B.key_a” column?

Answer 1

SELECT key_a, a.<<other columns>>, COUNT( b.key_b )
  FROM table_a a
       JOIN table_b b USING (key_a)
 GROUP BY key_a, a.<<other columns>>
HAVING COUNT( b.key_b ) > 1

will give you the data from table_a where there are multiple matching rows in table_b. If you are only interested in getting the key_a values that have multiple rows in table_b rather than getting all the data from table_a, you can remove the reference to table_a entirely and just do the GROUP BY and HAVING on table_b.