Question

I have table X, and each record in X has many records in Y.
I want a simple SQL which brings just the records from X which has the same set of records from Y.
Each group of records from X I want it to be distinguished from other group (which has another set of records from Y) with some value.

Example:

1 - 5
  - 6
2 - 3
3 - 5
  - 6
4 - 3
5 - 7

I want it back as:

1 - a
3 - a
2 - b
4 - b
5 - c

Answer 1

This looks like a job for wm_concat or stragg, which will give you your unique values without you having to transliterate them to something.

select x.id, stragg(x.value)
  from table_x x
 group by x.id

And with your comment you can guarantee order in a sub-query:

select id, substr(stragg(value),1,10)
  from ( select id, value
           from table_x
          order by id, value )
 group by id

The only problem is your values won’t be “pretty” they’ll be a concatenation of all the possible values you have in table_x

SQL> create table table_x ( id number, value number);

Table created.

SQL> insert into table_x values (1,5);

1 row created.

SQL> insert into table_x values (1,6);

1 row created.

SQL> insert into table_x values (2,3);

1 row created.

SQL> insert into table_x values (3,6);

1 row created.

SQL> insert into table_x values (3,5);

1 row created.

SQL> insert into table_x values (4,3);

1 row created.

SQL> insert into table_x values (5,7);

1 row created.

SQL> commit;

Commit complete.

SQL>
SQL> select id, substr(stragg(value),1,10)
  2    from ( select id, value
  3             from table_x
  4            order by id, value )
  5   group by id
  6         ;

        ID STRAGG(VALUE)
---------- ----------
         1 5;6
         2 3
         3 5;6
         4 3
         5 7

SQL>