Setting a composite key where the 2nd PK is generated depending on the 1st

Question

Hi,
I am dealing with a large database with a composite primary key made up of 1. ID 2.Sequence. Now my problem is for a particular ID I want to assign Sequence numbers starting from one to however many instances of that ID is there. This numbering will restart for another ID.Thus the combination of the ID and Sequence will make it unique.E.g.:
ID 1 has sequence 1,
ID 1 has sequence 2,
ID 1 has sequence 3,
ID 2 has sequence 1,
ID 2 has sequence 2....
For performance issues I will be using bulk bindings. Here is my go at the code. Can someone please suggest how I can optimize the code or make it better.
Code:
DECLARE
CURSOR c_orders IS
SELECT ID
FROM customer_orders;
TYPE t_num_array IS TABLE OF NUMBER INDEX BY BINARY_INTEGER;
v_order_ids t_num_array;
v_orders_sequence t_num_array;
v_row_count NUMBER := 0;
v_orders_sequence(1) := 1;
v_orders_sequence(2):=1;
BEGIN
OPEN c_orders;
LOOP
FETCH c_orders
BULK COLLECT INTO v_order_ids
LIMIT 100;
EXIT WHEN v_row_count = c_orders%ROWCOUNT;
v_row_count := c_orders%ROWCOUNT;
FOR i IN 2..v_order_ids.count LOOP
if(v_orders_id(i) = v_orders_id(i-1))
v_orders_sequence(i)=v_orders_sequence(i) + 1;
else
v_orders_sequence(i) = 1;
END LOOP;
FORALL i IN 1..v_order_ids.count
UPDATE customer_orders
SET sequence = v_orders_sequence(i)
WHERE id = v_order_ids(i);
END LOOP;
CLOSE c_orders;
COMMIT;
END;
Again if I made any mistakes please point it out and suggest any improvements to the code.
Edited by: user13579621 on Mar 3, 2011 3:40 PM
Edited by: user13579621 on Mar 3, 2011 3:43 PM
Edited by: user13579621 on Mar 4, 2011 7:15 AM
Edited by: user13579621 on Mar 4, 2011 7:19 AM
Edited by: user13579621 on Mar 4, 2011 9:35 AM

Answer 1

Hi,
The thing is both the id and the seqnums are part of a primary key. So say an ID has 3 instances this week and four the next week.
Week 1:
ID1 has Seq Num1
ID1 has Seq_num2
ID1 has Seq_num3
Week2:
ID1 has Seq Num1
ID1 has Seq_num2
ID1 has Seq_num3
ID1 has Seq_num4
Using an analytical function the sequence numbers will now be changed to reflect the new numbers. But what if the new ID coming in this week has a lesser ROWID in the original table. From what I understand your logic will now assign this new ID a sequence number of 1 and change the sequence numbers of the 3 existing from last week from 1 to 2,3,4 accordingly. But since it is part of the primary key we dont want to change them. The new ID coming in should always be assigned a sequence number of 4 not 1. Any ideas about how to achieve this.
Also it does not have to be the way I do it. Let me just go back and rephrase what I want to achieve.
I have a new table primary key(id,seqnum) which gets data from another table. Now I want to set the id incremented normally and the sequence number incremented uniquely like I mentioned before for each unique id and then reset for the next ID. How do I go about achieving this efficiently, keeping in mind that if a new id comes in next week the new id will always be assigned a sequence number that is +1 of the last sequence number for that ID?