How to find leastly used non-clustered indexes in SQL Server ?

  • Article

Script:

 

SELECT objectname=OBJECT_NAME(s.OBJECT_ID)

, indexname=i.name

, i.index_id 

, reads=user_seeks + user_scans + user_lookups 

, writes = user_updates 

, p.rows

FROM sys.dm_db_index_usage_stats s JOIN sys.indexes i

ON i.index_id = s.index_id AND s.OBJECT_ID = i.OBJECT_ID 

JOIN sys.partitions p ON p.index_id = s.index_id AND s.OBJECT_ID = p.OBJECT_ID

WHERE OBJECTPROPERTY(s.OBJECT_ID,'IsUserTable') = 1 

AND s.database_id = DB_ID() �

AND i.type_desc = 'nonclustered'

AND i.is_primary_key = 0

AND i.is_unique_constraint = 0

AND p.rows > 10000 AND(user_seeks + user_scans + user_lookups)< user_updates ORDER BY reads, rows DESC

 

It is basically finding the index whose user_updates(writes) are more than reads.