Question

I’m trying to write a recursive function of sorts in PL/SQL.

The problem is:

So say table A has rows:

 {B, C},
 {C, D},
 {C, F},
 {D, E},
 {E, F}

Return everything that B is dependent on, directly and indirectly.

The tuple {B, C} implies that B is dependent on C, C is dependent on D and so on and so forth.
This function, when given B, would return a cursor or something that would yield: {C, D, F, E}
Notice that simple looping through and just printing values may yield duplicate results (in this case, E).

I’m rather new to PL/SQL and I can’t really think of a way to do this.

Thanks in advance for any help!

Answer 1

Assuming the table looks like this:

ID    PARENT_ID
---   ---------
B     C
C     D
C     F
D     E
E     F

Why wouldn’t you use a hierarchical query like:

select distinct parent_id
  from (select parent_id
          from my_table
         start with ID = 'B'
               connect by nocycle id = prior parent_id
       )
 order by parent_id

This SQL’s untested, but it should be a point in the right direction; your function could return that cursor if that’s what you needed, or an array of values if not.