March 31, 2008 at 7:44 am
I have three large tables where I need to completely restructure the indexes, including changing the clustering index. The table sizes are as follows:
1. 157 GB data, 97 million rows.
2. 24 GB data, 775 million rows.
3. 19 GB data, 68 million rows.
After a bit of trial and error, I am leaning towards unloading the tables using BCP. Truncating the tables. Changing the indexes and reloading the data.
Is that the fastest way.
March 31, 2008 at 2:19 pm
I'd drop the indexes before truncating and recreating them AFTER loading the data. It will allow the reload to occur that much quicker because it won't be hitting your indexes.
-- You can't be late until you show up.
March 31, 2008 at 2:23 pm
I can see dropping the non-clustered indexes before reloading the table. What about the clustered index?
Also, does it help to have the input file sorted in cluster index order?
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply