Nested for loop in the collections

Question

Hi Experts,
collection1
============
SELECT o.object_id
      BULK COLLECT INTO l_obj_info
        FROM (SELECT     n.node_id, n.object_id
                    FROM nodes n
              START WITH n.node_id = 100
              CONNECT BY PRIOR n.node_id = n.parent_node_id) n
             INNER JOIN
             objects o ON n.object_id = o.object_id
       WHERE o.object_type_id = 285;
collection2
============
SELECT *
      BULK COLLECT INTO l_tab
        FROM ((SELECT     REGEXP_SUBSTR (i_l_text, '[^,]+', 1, LEVEL)
                     FROM DUAL
               CONNECT BY REGEXP_SUBSTR (i_l_text, '[^,]+', 1, LEVEL) IS NOT NULL));
   END;
collection3
============
SELECT o.object_id
               BULK COLLECT INTO l_fin_tab
                 FROM objects o JOIN ATTRIBUTES att
                      ON o.object_id = att.object_id
                WHERE o.object_id = collection1.object_id
                  --AND att.VALUE = collection2.val;
Please tell me how to implement for loop in the collection3 to get the values from collection1 and collection2.
i have tried in the below way
CREATE OR REPLACE TYPE LIST_OF_ATTRIBUTES_TYPE AS TABLE OF varchar2(4000);
CREATE OR REPLACE TYPE LIST_OF_OBJECT_IDS_TYPE AS TABLE OF number(9);
CREATE OR REPLACE FUNCTION f_get_objects_by_type_id (
   i_object_type_id   IN   NUMBER,
   i_l_text           IN   VARCHAR2,
   i_scope_node_id         NUMBER
   RETURN list_of_object_ids_type
AS
   CURSOR objs_info
   IS
      SELECT o.object_id
        FROM (SELECT     n.node_id, n.object_id
                    FROM nodes n
              START WITH n.node_id = i_scope_node_id
              CONNECT BY PRIOR n.node_id = n.parent_node_id) n
             INNER JOIN
             objects o ON n.object_id = o.object_id
       WHERE o.object_type_id = i_object_type_id;
   l_tab       list_of_attributes_type := list_of_attributes_type ();
   --l_obj_info   list_of_object_ids_type := list_of_object_ids_type ();
   l_fin_tab   list_of_object_ids_type := list_of_object_ids_type ();
BEGIN
   BEGIN
      SELECT *
      BULK COLLECT INTO l_tab
        FROM ((SELECT     trREGEXP_SUBSTR (i_l_text, '[^,]+', 1, LEVEL)
                     FROM DUAL
               CONNECT BY REGEXP_SUBSTR (i_l_text, '[^,]+', 1, LEVEL) IS NOT NULL));
   END;
   IF l_tab.COUNT > 0
   THEN
      FOR i IN objs_info
      LOOP
         FOR j IN l_tab.FIRST .. l_tab.LAST
         LOOP
            SELECT o.object_id
            BULK COLLECT INTO l_fin_tab
              FROM objects o JOIN ATTRIBUTES att ON o.object_id =
                                                                 att.object_id
             WHERE
                            att.VALUE = l_tab (j) and o.object_id =objs_info(i);
         END LOOP;
      END LOOP;
   END IF;
   RETURN l_fin_tab;
END f_get_objects_by_type_id;

Answer 1

Why are you wanting to do this?
It looks like you are trying to implement SQL joins in PL code.  Not only is that using up expensive PGA memory by storing the data in collections, but doing such retrieval of data to try and join it in PL loops, is never going to be as fast as just joining the SQL queries using SQL itself.
Post some example data and your database version, with an example of what the output should look like from that example data.
Re: 2. How do I ask a question on the forums?