Question

I am trying to write a SQL script to fetch data for a specific date range. In the example shown below I am trying to pull data for 1 day (datetime) only.

But it just take more than 3 mins to complete the query execution.

I am not quite sure if I am missing any indexes to create on Datetime column. If yes could anybody please suggest me how to make this SQL query a quick execution?

Please also see the snapshots attached from query execution plan and statistics IO.

Thank you very much for your help.

The SQL query

set Statistics IO on
declare @StartDate datetime,@EndDate datetime
set @StartDate = '2012-07-19 00:00:00.000'
set @EndDate = '2012-07-20 23:59:00.000'
select * from Admin_Letters A with (nolock)
Where A.Date_Linked > @StartDate and A.Date_Linked < @EndDate

The Execution plan snapshot(s)

1.

2.

Answer 1

With a SELECT * in your query, you will not have much luck trying to optimize anything…. since you request all columns from the table, SQL Server’s query optimizer will often go for a table scan since that’s just quicker.

Also: table scan implies that you don’t have any clustering index on your table –> horribly bad choice. You should have a good clustering index on every real table!

Once you have that (on e.g. an INT IDENTITY column), then you need to decide which columns you really need from that table. Put a nonclustered index on Date_Linked, and include all those columns you really truly need – something like:

CREATE NONCLUSTERED INDEX IX01_Admin_Letters
ON dbo.Admin_Letters(Date_Linked)
INCLUDE (Col1, Col2, ..., ColN)   -- as **few** columns as possible!

and then a

SELECT Date_Linked, Col1, Col2, ...ColN 
FROM dbo.Admin_Letters A
WHERE A.Date_Linked > @StartDate AND A.Date_Linked < @EndDate

will really fly! (because now, SQL Server’s query optimizer can choose to just pick a few rows from the index – and even an index scan, on a few columns, will be lots faster than a table scan)