In my database there are around 1500 tables and mostly on all the tables there are non-unique indexes created.I want to know how these non-unique indexes will help improve my performance and how I can drop unwanted indexes which are occupying space in my database.
Determine if the indexes are selective or non-selective, based on the ratio of distinct keys to the total # keys in the index and if they are not used for specific queries, they you can drop them. It might be a good idea to have atleast 1 unique indes per table.
Since u have so many of them, find out why they were there in the first place, generally it is created for for Fast Full scans of Index in Data Warehouse kind of scenario.
Also it is used to bypass parent-child relationship during initial inserts.
Also may be it is true what Halo has said.
So you have to find out why?
Click Here to Expand Forum to Full Width