Bulk Update query

Question

Hello Friends,
Can some one suggest me is that any possible bulk update query which consumes less timings?
Table - MyTable
id - PK.
orderid - Order Id.
Subid - Sub Id for an Order.
lineitem - LineItemId.
ProducId - Product Id.
Now i want to update the Subid to My Table for Every Order on Basis of LineItemId..
For Ex:
I will be having the records in MyTable as for a single Order...there can mutilple Subid's.
UPDATE MyTable SET subid = 1 WHERE orderid = 123 AND lineitem = 1;
UPDATE MyTable SET subid = 1 WHERE orderid = 123 AND lineitem = 2;
UPDATE MyTable SET subid = 5 WHERE orderid = 123 AND lineitem = 2000;
I worked out three scenarios as follows,
Case1:
UPDATE MyTable SET subid = 5 WHERE orderid = 123 AND lineitem = 2000;
Case2:
UPDATE MyTable SET subid = 1 WHERE orderid = 123 AND lineitem in(1,2,3.....1000);
UPDATE MyTable SET subid = 2 WHERE orderid = 123 AND lineitem in(1001,1002,.....1100);
Case3:
UPDATE MyTable SET subid= CASE WHEN lineitem = 1 THEN 1 WHEN lineitem = 2 THEN 2 .....WHEN 1000 THEN 1000 END WHERE orderid = 123;
Please suggest me which update consumes less time and helpful for updating more records nearly 5000 - 10000 at a single table.

Answer 1

You are comparing three cases that are not equal to each other:
Case1:
UPDATE MyTable SET subid = 5 WHERE orderid = 123 AND lineitem = 2000;
Here you update the records with orderid = 123 and lineitem = 2000
Case2:
UPDATE MyTable SET subid = 1 WHERE orderid = 123 AND lineitem in(1,2,3.....1000);
UPDATE MyTable SET subid = 2 WHERE orderid = 123 AND lineitem in(1001,1002,.....1100);
This are multiple update statement to update all records with orderid = 123 and lineitems between 1 and 1100.
Case3:
UPDATE MyTable SET subid= CASE WHEN lineitem = 1 THEN 1 WHEN lineitem = 2 THEN 2 .....WHEN 1000 THEN 1000 END WHERE orderid = 123;
And here all records with orderid = 123, regardless of the lineitem are updated.
So my guess is that 1 will be the fastest as it is updating the least amount of records, followed by 2 and then 3. But it is a really weird comparison.
I think you'd better make up your mind first about which records need to be updated and how. And then it is best to use one update statement to do the job.
Regards,
Rob.