November 13, 2003 at 11:31 am
Is the SORT_IN_TEMPDB option available when I create a Primary Key Clustered
constraint? This option is available when I do the same (ie Clustered Unique)
as an index but erros when I do through constraint..
Here is the background, I must enforce the uniqueness thru the PK, since I have these
tables as part of a partitioned view( and it requires the partitioned columns
to be part of the PK). I am recreating the constraint since at the end of
each quarter I drop the corresponding partition table and recreate the constraint/index
so I have 100% contiguousness. I would like to use SORT_IN_TEMPDB
since that makes them more contiguous than doing in the user database itself.
Any ideas?
thanks.
November 13, 2003 at 12:54 pm
After creating them, you can use this on PKs:
CREATE UNIQUE CLUSTERED INDEX pk_Table ON Table(PKCol) WITH DROP_EXISTING, SORT_IN_TEMPDB
--Jonathan
--Jonathan
November 13, 2003 at 3:04 pm
That works great. Thanks you.
It is interesting I needed to run it a couple of times to get it to 0% fragmentation.
I would prefer to get the old partitions as close to contiguous as possible, since no
data is loaded into them after the quarter is past.
Thanks Again.
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply