Numeric overflow error using binary integer

Question

Hi experts,
I am facing issue while solving a numeric overflow error. after analyzing we came to know that in the below code BINARY_INTEGER is causing the issue as input is exceeding its range. I tried to replace BINARY_INTEGER by varchar2(20) but its saying
"Error(580,20): PLS-00657: Implementation restriction: bulk SQL with associative arrays with VARCHAR2 key is not supported."
We need to remove this binary_integer. I dont know how to do this. Can anybody give some idea or what code change required here ? thanks in advance. Cheers.. Below is the code,
===================================================
   PROCEDURE UpdateCost_
      p_Cost_typ IN OUT NOCOPY CM_t,
   IS
      TYPE ObjektIdTab_itabt IS TABLE OF ObjektId_tabt INDEX BY BINARY_INTEGER;
      v_cost_IdTab_itab ObjektIdTab_itabt;
      v_CM_ID INTEGER := p_Cost_typ.costm.CM_ID;
      BEGIN
            SELECT CAST(MULTISET
                    (SELECT Costwps.CMKostId
                      FROM CM_Pos_r NRPos,
                            CMK_z_r costzpps,
                            CMG_Cost_v Costwps
                      WHERE NRPos.CM_ID = v_CM_ID
                        AND NRPos.SNRId_G = SNRCT.SNRPos.SNRId_G
                        AND costzpps.CM_ID = NRPos.CM_ID
                        AND costzpps.CMSNRPosId = NRPos.CMSNRPosId
                        AND costzpps.Kost_s = Kost.Costnzl.Kost_s
                        AND Costwps.CMKz_Id = costzpps.CMKz_Id
                        AND Costwps.TypCode NOT IN
                            (SELECT kw.TypCode
                               FROM TABLE(Kost.Kostwt_tab) kw
                    ) AS ObjektId_tabt )
              BULK COLLECT
              INTO v_cost_IdTab_itab
              FROM TABLE(p_Cost_typ.SNR_tab) SNRCT,
                   TABLE(SNRCT.Kost_tab) Kost
         FOR v_i IN 1 .. v_cost_IdTab_itab.COUNT LOOP
            FOR v_j IN 1 .. v_cost_IdTab_itab(v_i).COUNT LOOP
               DELETE FROM CMG_Cost_v WHERE CMKostId = v_cost_IdTab_itab(v_i)(v_j);
            END LOOP;
         END LOOP;
END;
===================================================

Answer 1

Thanks for your reply. I tried with INDEX by NUMBER. but oracle says its not a valid use of index by thing. and moreover I also tried with by removing INDEX BY clause. but in that case we are not at all getting any data in for loop. some people says to use extend clause. But again I am not sure How to do so. Can you please let me know code for this.
I know you are trying to help by you need to STOP telling us what problem you have and SHOW US. Saying 'Oracle says' is useless. Post EXACTLY what code you are using, the EXACT steps you are using to compile that code and the EXACT result that you are getting.
You also made no comment about the 'overflow' issue. A BINARY_INTEGER (PLS_INTEGER) has a very large range of values:
http://docs.oracle.com/cd/E11882_01/appdev.112/e25519/datatypes.htm#i10726
>
The PLS_INTEGER data type stores signed integers in the range -2,147,483,648 through 2,147,483,647, represented in 32 bits.
>
If you are trying to create a collection of more than 2 BILLION of anything you have a serious problem with either WHAT you are trying to do or HOW you are trying to do it. Your 'overflow' issue is more likely a symptom that you are really running out of memory. You should ALWAYS have a LIMIT clause when you do BULK COLLECT statements.
Also see this section in that doc: SIMPLE_INTEGER Subtype of PLS_INTEGER
You need to address your LIMIT issue first and then address any other issues that arise from actually executing the code.
Then see the section 'SELECT INTO Statement with BULK COLLECT Clause' in that doc
http://docs.oracle.com/cd/E11882_01/appdev.112/e25519/tuning.htm#BABEIACI
That section has an example that shows you do NOT need to use an INDEX BY clause to create collections as you are trying to do. So your not 'getting any data in for loop' is NOT related to the lack of that clause.
That example also shows you that you do NOT use 'extends' when doing BULK COLLECT. The bulk collection automatically extends the collection as needed to hold the entire results (assuming you don't run out of memory for 2 BILLION things).
Example 12-22 in that same doc shows the proper way to use a double loop and a BULK COLLECT with a LIMIT clause
http://docs.oracle.com/cd/E11882_01/appdev.112/e25519/tuning.htm#BABCCJCB
Here is very simple sample code you can use in the SCOTT schema to understand how the double loop and LIMIT clauses work together.
>
The FETCH does a BULK COLLECT of all data into 'v'. It will either get all the data or none if there isn't any.
The LOOP construct would be used when you have a LIMIT clause so that Oracle would 'loop' back to
get the next set of records. Run this example in the SCOTT schema and you will see how the LIMIT clause works.
I have 14 records in my EMP table.
DECLARE
  CURSOR c1 IS (SELECT * FROM emp);
  TYPE typ_tbl IS TABLE OF c1%rowtype;
  v typ_tbl;
BEGIN
  OPEN c1;
  LOOP                                                 --Loop added
    FETCH c1 BULK COLLECT INTO v LIMIT 3; -- process 3 records at a time
        -- process the first 3 max records
       DBMS_OUTPUT.PUT_LINE('Processing ' || v.COUNT || ' records.');
        FOR i IN v.first..v.last LOOP
            DBMS_OUTPUT.PUT_LINE(v(i).empno);
        END LOOP; 
    EXIT WHEN c1%NOTFOUND;
  END LOOP;
  DBMS_OUTPUT.PUT_LINE('All done');
END;
In the FOR loop you would do any processing of the nested table you want to do
and could use a FORALL to do an INSERT into another table.
>
I strongly suggest that you modify your code to work with a VERY SMALL set of data until it works properly. Then expand it to work with all of the data needed, preferably by using an appropriate LIMIT clause of no more than 1000.