SQL Server – Finding Parent Source from T-SQL Snippet

All we need is an easy explanation of the problem, so here it is.

The environments for which I’m responsible have some pretty substantial plan non-reuse challenges. I’ve run across the following query (h/t Brent Ozar blog commenter Michael J Swart) that does a fine job of itemizing the worst offenders:

WITH cte AS (
   SELECT COUNT(*) [count], query_hash, min(sql_handle) [sql_handle_example]
   FROM sys.dm_exec_query_stats
   GROUP BY query_hash
   )
SELECT cte.*, t.text [query_text_example]
FROM cte
CROSS APPLY sys.dm_exec_sql_text(sql_handle_example) t
WHERE [count] > 100
ORDER BY [count] DESC

My challenge is taking a snip of the [query_text_example] text and efficiently identifying whether it’s originating from a sproc, and if so which one in which database. I’ve done some Googling and testing and it’s been puzzlingly difficult to find a solution that takes a snippet of query text, whether it was dynamically built or not, and accurately identifies its parent sproc. Does anyone have any suggestions?

How to solve :

I know you bored from this bug, So we are here to help you! Take a deep breath and look at the explanation of your problem. We have many solutions to this problem, But we recommend you to use the first method because it is tested & true method that will 100% work for you.

Method 1

parental advisory

You can use the dm_exec_procedure_stats view, and if you’re on SQL Server 2016 or better, the dm_exec_function_stats view as well, to track down scalar UDFs.

Both of those views have an object_id column that can be used to resolve procedure and function names with the OBJECT_NAME function. You can match those to various other views on the sql_handle or plan_handle column.

Unless you are doing this scoped to a single database that you care about, you may also need to use the dm_exec_plan_attributes view to grab the dbid attribute, which can be used as a second argument in OBJECT_NAME.

Method 2

dm_exec_sql_text supplies a objectid column which is NULL for ad-hoc batches. You can just left-join it with sys.objects

WITH cte AS (
    SELECT
      COUNT(*) [count],
      s.query_hash,
      min(s.sql_handle) [sql_handle_example]
    FROM sys.dm_exec_query_stats s
    GROUP BY s.query_hash
    HAVING count(*) > 100
)
SELECT
  cte.*,
  t.text [query_text_example],
  o.name
FROM cte
CROSS APPLY sys.dm_exec_sql_text(cte.sql_handle_example) t
LEFT JOIN sys.objects o ON o.object_id = t.objectid
ORDER BY [count] DESC;

This only works for a single database. To join it across multiple unknown databases you need dynamic SQL. Each database has a separate sys.objects.

Note: Use and implement method 1 because this method fully tested our system.
Thank you 🙂

All methods was sourced from stackoverflow.com or stackexchange.com, is licensed under cc by-sa 2.5, cc by-sa 3.0 and cc by-sa 4.0

Leave a Reply