Join Logic Help

Question

I'm having difficulty with the join logic of a particular query, which is modeled with the following (highly simplified) schema:
create table a(
    a_id number,
    constraint pk_a primary key(a_id));
create table b(
    b_id number,
    a_id number,
    constraint pk_b primary key(b_id),
    constraint fk_b_a foreign key(a_id) references a(a_id));
create table c(
    c_id number,
    b_id number,
    constraint pk_c primary key(c_id),
    constraint fk_c_b foreign key(b_id) references b(b_id));
insert into a values(1);
insert into a values(2);
insert into b values(1,1);
insert into b values(2,1);
insert into b values(3,2);
insert into b values(4,2);
insert into c values(1,1);
insert into c values(2,2);
insert into c values(3,3);There can be 1-many "b" records associated with each "a" record, but "b" and "c" have a 1-1 relationship. What query will select only the "a" records that have ALL associated "b" records contained in "c"? Given the above data, the query should return only "a=1", since both "b=1" and "b=2" are contained in "c". However, it would NOT return "a=2", since both "b=3" and "b=4" are not contained in "c". The results should be:
OUTPUT:
    a_id
    1

Answer 1

Hi,
Here's one way:
SELECT       a.a_id
FROM           a
JOIN           b  ON  a.a_id     = b.a_id
LEFT OUTER JOIN      c  ON     b.b_id     = c.b_id
GROUP BY  a.a_id
HAVING       COUNT (b.b_id)  = COUNT (c.b_id)
;Most things that you can do with IN sub-queries can also be done with joins, or with EXISTS sub-queries. (Most things that can be done with EXISTS sub-queries can also be done wioth joins or IN sub-queries. However, there are lots of things that can be done only with joins.)
If you don't like one, try the others.
Thanks for posting the CREATE TABLE and INSERT statements; that's very helpful.
Don't forget to post the results you want from that data, no matter how simple those results are, and your version of Oracle.