The below written query helps us find the most expensive queries from a read IO perspective. The DMV that we have used in this script is sys.dm_exec_query_stats and the function that we have used is sys.dm_exec_sql_text.
SELECT TOP 5 qt.text AS 'SQL', qstats.total_physical_reads AS 'Total Physical Reads',
qstats.total_physical_reads/qstats.execution_count AS 'Average Physical Reads',
qstats.execution_count AS 'Execution Count',
qstats.total_worker_time/qstats.execution_count AS 'Average Worker Time',
qstats.total_worker_time AS 'Total Worker Time',
DATEDIFF(Hour, qstats.creation_time, GetDate()) AS 'AgeInCache In Hours', db_name(qt.dbid) AS 'Database Name'
FROM sys.dm_exec_query_stats AS qstats
CROSS APPLY sys.dm_exec_sql_text(qstats.sql_handle) AS qt
WHERE qt.dbid = db_id() -- Filter by current database
ORDER BY qstats.total_physical_reads DESC 
Top 5 expensive Queries from a Write IO perspective
2013-10-03 (first published: 2013-09-09)
2,334 reads
With SQL Server now supported on Linux, traditional SQL Server DBAs will need to learn how to work with Linux, and Linux administrators will need to learn how to work with SQL Server. Fortunately, there are several tools available to assist. In this article Prashanth Jayaram describes how to work with the DMV tool which treats queries as files.
2018-04-09
2,500 reads
The SQL Server 2016 Query Store provides several new ways of troubleshooting queries, studying their plans, exploring their context settings, and checking their performance metrics. In using the Query Store to ensure that performance is as good as possible, it isn't long before it becomes important to be familiar with the DMVs that are associated with the query store, and using them in custom queries.
2016-02-09
3,612 reads
2015-09-01
1,444 reads
2015-08-27
1,265 reads