|
|
|
Say Hey Kid
      
Group: General Forum Members
Last Login: Wednesday, May 01, 2013 5:17 PM
Points: 701,
Visits: 207
|
|
|
|
|
|
Old Hand
      
Group: General Forum Members
Last Login: Friday, May 17, 2013 8:38 AM
Points: 377,
Visits: 360
|
|
I am not sure I understand what is considered a duplicate index. It appears that if more than one index starts with the same field, that it is considered a duplicate index. I have 4 indexes on one table that all have the same field as the first member. Why are these considered duplicate indexes?
Thanks Glen
|
|
|
|
|
Say Hey Kid
      
Group: General Forum Members
Last Login: Wednesday, May 01, 2013 5:17 PM
Points: 701,
Visits: 207
|
|
It only outputs indexes that have common first seek columns. Indexes that have common first seek columns are dupes or might be merge-able, or the one of the indexes might be able to get their seek columns rearranged so you can have more index usage possibilities without adding more indexes.
I would only intentially have indexes with common first seek columns on tables with hardly any write activity.
The only other reason to dupe on first seek columns is to dupe on PKs that experience a lot of multi-row up-stream (from many to one; Where is on the many table) joins that cause bookmark lookups. PKs (also UQ constraints) can't have include columns, so you have to create another nonclustered index that has the seem seek as the PK and then add include columns.
|
|
|
|