Nice explanation!

We Can also get the exact duplicate index using below query.

with indexcols as
(
select object_id as id, index_id as indid, name,
(select case keyno when 0 then NULL else colid end as [data()]
from sys.sysindexkeys as k
where k.id = i.object_id
and k.indid = i.index_id
order by keyno, colid
for xml path(”)) as cols,
(select case keyno when 0 then colid else NULL end as [data()]
from sys.sysindexkeys as k
where k.id = i.object_id
and k.indid = i.index_id
order by colid
for xml path(”)) as inc
from sys.indexes as i
)
select
object_schema_name(c1.id) + ‘.’ + object_name(c1.id) as ‘table’,
c1.name as ‘index’,
c2.name as ‘exactduplicate’
from indexcols as c1
join indexcols as c2
on c1.id = c2.id
and c1.indid < c2.indid
and c1.cols = c2.cols
and c1.inc = c2.inc;

– Script to get Overlapping Indexes
with indexcols as
(
select object_id as id, index_id as indid, name,
(select case keyno when 0 then NULL else colid end as [data()]
from sys.sysindexkeys as k
where k.id = i.object_id
and k.indid = i.index_id
order by keyno, colid
for xml path('')) as cols
from sys.indexes as i
)
select
object_schema_name(c1.id) + '.' + object_name(c1.id) as 'table',
c1.name as 'index',
c2.name as 'partialduplicate'
from indexcols as c1
join indexcols as c2
on c1.id = c2.id
and c1.indid < c2.indid
and (c1.cols like c2.cols + '%'
or c2.cols like c1.cols + '%') ;

Hope, it would be really help you to get exact/Overlapping duplicate indexes in any Database.
Note:- Please double check these scripts on Development server before executing it on Production server.

Feel free to redirect me in case of any query/questions.

Regards,
Girijesh

Jason

Kim Tripp has expended on index duplicate searching & given her own script at: http://sqlskills.com/BLOGS/KIMBERLY/post/UnderstandingDuplicateIndexes.aspx and http://sqlskills.com/BLOGS/KIMBERLY/post/RemovingDuplicateIndexes.aspx

Dave

Nice article. I tried to run this on my DEV instance and found this interesting thing. I am having same named table with same named primary key column in different Schema. This script display indexes on these columns as duplicates. I believe its not correct.

Please share you thoughts

Results:
SchemaName | TableName | IndexName | OverLappingIndex | Col1
=========================================================
E2S | queue | PK_queue | PK_email | Id
Email | Queue | PK_email | PK_queue | Id

I’m interested in more information on this comment: “Left-based subsets the same doesn’t mean you can drop the wider one “. Using the example of two indexes (colA,colB,colC) and (colA,colB), assuming that everything else in the index is equal (includes, filters, etc) and that ColC is not very wide.

Can we say the index 2 is redundant because index 1 includes the same cols as index 2?

I’d like to know your ‘rule-of-thumb’ approach to this scenario as it’s fairly common.

Thanks,
Dave

I totally agree. Collecting input to modify this query to reflect the all the suggestions. Your suggestion is very valid and based your your suggestion I will further update the same.