SQL Query Tuning

I've following table and data

tblSEL
Date_Taken | Time | Main_ID | Value
-----------------------------------------------
5/11/2008 12:00:00 AM | 500 | 233 | 4780
5/11/2008 12:00:00 AM | 500 | 239 | 3122
5/11/2008 12:00:00 AM | 515 | 233 | 5435
5/11/2008 12:00:00 AM | 530 | 233 | 6799
5/11/2008 12:00:00 AM | 545 | 233 | 4090
5/11/2008 12:00:00 AM | 600 | 234 | 2312
5/11/2008 12:00:00 AM | 615 | 233 | 7409
...
...
...
tblSEL having more than 6million rows
*Date_Taken datatype is smalldatetime

tblLocation
Main_ID | Location
---------------------
233 | PRK
234 | PER
239 | SAB
...
...
...

i've following query to getting the last 7month data
SELECT
t1.Date_Taken, t1.Time,
t1.Main_ID, t1.WATER_ULEVEL
FROM dbo.tblSEL t1 INNER JOIN dbo.tblLocation t2
ON t1.Main_ID=t2.Main_ID
WHERE t2.Location='PRK'
AND t1.Date_Taken>=CONVERT(VARCHAR(10),DATEADD(m,-7,GETDATE()),101) AND
t1.Date_Taken<=CONVERT(VARCHAR(10), GETDATE(), 101)
ORDER BY t1.Date_Taken, t1.Time

How to adjust this query to make sure i get same data and also run at the best performance.

AND t1.Date_Taken<=CONVERT(VARCHAR(10), GETDATE(), 101)
Why is the line above necessary in the query ? Does tblSel ever contain Date_Taken values that are higher than GETDATE() ? If the latest Date_Marker values are always today's date then you can leave out this additonal clause to optimize the query slightly.
As for indexing; I will assume that tblLocation already has a clustered index on Main_ID as the Primary Key of that table. Placing a nonclustered index on Main_ID in tblSel will help if you have not already done so. Placing an index on the Location column in tblLocation might help too, but if tblLocation contains less than 1000 records then the performance advantage will only be small.

Hi
If t1.Date_Taken is VARCHAR(10) and you are using style 101 (mm/dd/yyyy) your query will be return not only the last 7 months data, but the previous years data during the same period. See example here.create
table #test1(Date_Taken
VARCHAR(10))INSERT
INTO #test1 (Date_Taken) SELECT '05/22/2007'INSERT
INTO #test1 (Date_Taken) SELECT '05/22/2008'INSERT
INTO #test1 (Date_Taken) SELECT '05/23/2008'INSERT
INTO #test1 (Date_Taken) SELECT '05/24/2007'select
* from #test1 where date_taken >= '05/21/2008' AND date_taken <= '05/24/2008'
Now if it is DATETIME and you are trying to cut off the time then following example does work. Notice I work out the date range before using them. This will increase performance if you are using a stored procedure to do the selection. Now you can pass in the -7 in to the stored procedure. I take it the Date_Taken field is indexed. Is it clustered?create table #test1(
Date_Taken
DATETIME)INSERT
INTO #test1 (Date_Taken) SELECT '11/22/2007 12:00:00'INSERT
INTO #test1 (Date_Taken) SELECT '05/22/2008 17:00:00'INSERT
INTO #test1 (Date_Taken) SELECT '05/23/2008 15:00:00'INSERT
INTO #test1 (Date_Taken) SELECT '11/24/2006 17:00:00'DECLARE @Date1 DATETIME
DECLARE
@Date2 DATETIMESET
@Date1 = CONVERT(DATETIME, CONVERT(VARCHAR(10), GETUTCDATE(), 101))SET
@Date1 = DATEADD(m, -7, @Date1)SET
@Date2 = CONVERT(DATETIME, CONVERT(VARCHAR(10), GETUTCDATE(), 101) + ' 23:59:59')SELECT
@Date1, @Date2select
* from #test1 where date_taken BETWEEN @Date1 AND @Date2
I hope this helps.
Regards
Richard...