all groups > sql server data warehouse > july 2003 >
You're in the

sql server data warehouse

group:

Drop pk clustered index on insert.


Drop pk clustered index on insert. Phil
7/29/2003 6:34:00 AM
sql server data warehouse:
I need to increase performance on a fact table insert.
There is one Primary Key identity column defined on
the fact table with a clustered index defined by default.
Can I drop the constraint and the index on this column
or maybe just the clustered index to increase
performance for inserted rows. Any suggestions ?
Re: Drop pk clustered index on insert. Jacco Schalkwijk
7/29/2003 2:56:51 PM
You can't drop the index without dropping the Primary Key, but because the
clustered index is on an identity column I don't think you will see much
improvement in performance in insertions when you drop the key (index)
first, and recreate it afterwards. Doing a sort on a large table, which will
happen if you create a clustered index on it, can be very expensiv. If the
number of rows that is inserted is small compared to the number of rows that
is already in the table it might even perform worse. But you have to test
both situations, YMMV.

--
Jacco Schalkwijk MCDBA, MCSD, MCSE
Database Administrator
Eurostop Ltd.


[quoted text, click to view]

AddThis Social Bookmark Button