Great work!
Tim

An Index Scan is not the same as a table scan. A “CLUSTERED” index scan is the same as a table scan.

If you have sql server 2000 then
1. Why dont you run Index Tuning Wizard, analyze your query, result of which recommends any improvements can be done on indexes for that query.

If you have SQL Server 2005,
1. Use Database Engine Tuning Advisor. This does the same thing, it will give you recommendations about indexes for that particular query.

Regards,
IM.

I have a table with abt 200K records. a “where clause” on a non-indexed field causes a “clustered index scan”. It works fine and fast, but perf degrades when concurrent users increase.

do you think a clustered index scan can be a good reason for perf hit with increase in concurrent users even in a table of this size. OR maybe i should keep looking further for other causes.

rgds

Ajay

I should optimize a very simple query (select) where the actual execution plan presents a clustered index cost of 56 % due to estimated cost operator. how can i reduce this cost??

thanks

Here is my Select statement

UPDATE FCE
SET
–FCE.RECAT_CREDIT_CMT = RIID.CREDIT_CMT,
FCE.RECAT_DEAL_ID = RIID.DEAL_ID,
FCE.RECAT_STATUS = RIID.STATUS,
FCE.RECAT_STATUS_DT = RIID.CLM_STATUS_DT,
FCE.RECAT_EUV_FLAG_ITEM = RIID.EUV_LINE_FLAG,
FCE.RECAT_EUV_FLAG = RIID.EUV_FLAG,
FCE.M_CODE = RIID.MCODE,
FCE.RECAT_EUV_SUBMIT_DT =RIID.EUV_SUBMIT_DT,
FCE.RECAT_EUV_CONTROL_DT =RIID.EUV_CONTROL_DT,
FCE.STAKO_DEAL=RIID.STAKO_DEAL,
FCE.MMP_FLAG=RIID.MMP_FLAG
FROM FinClaimEntries FCE, RECAT_ITEMS RIID
where FCE.CLAIM_ID = RIID.ECLAIM_ID
AND (FCE.recat_prod_item_id = RIID.ECLM_ITEM_ID
OR FCE.recat_bdl_item_id = RIID.ECLM_ITEM_ID
OR FCE.CLAIM_ITEM_ID = RIID.ECLM_ITEM_ID
)
AND RIID.CLM_ID = (SELECT MAX(CLM_ID) FROM RECAT_ITEMS RIID2 where FCE.CLAIM_ID = RIID2.ECLAIM_ID )

Here is Number of Records Details

FinClaimEntries – 78,589,25
RECAT_ITEMS – 97,256,86

Please let me know yours suggestion for Index Creation & Joins

Thanks,
Sunil

Here’s the query i was talking about.. I tested it in SQL 2000 and SQL 2005 both…

DECLARE @StartDate DATETIME
DECLARE @EndDate DATETIME
SET @StartDate = ‘2008-09-11 16:00:00′
SET @EndDate = ‘2008-09-12 00:00:00′
SELECT * FROM PCHDotCom.dbo.BillMeEntries WITH (NOLOCK, INDEX=IX_BME_CREATED)
WHERE
Created >= @StartDate AND
Created < @EndDate
ORDER BY
1 DESC

If I remove the index hint, the query plan will decide to do clustered index scan and the query takes drastically longer

Please help ASAP.

Thanks,
Sunil

Your explainations are really amazing.

I have seen couple of times with SQL Server 2000 that for some reason, query optimizer do not use index on its own and query finishes in about 8 min while if I give hint(explicit) then it returns output in less than a min. I checked index frag and update stats.. all are looking fine.

Can you please advice ?

Thanks,
sunil

I am trying to implement index by using substring function but its throwing an error so could u suggest me for better solution.

Generally, implementing this substring function in index for getting the index seek on the query.

Below is the syntax.

create nonclustered index IX_CreatedBy on TransDetails
(substring(createdby, charindex(‘\’,CreatedBy) + 1, len(CreatedBy)))

in the db having the value like \
at this moment cannot change the table structure for this query alone.

Regards,
Mohanraj Jayaraman

Table scan (called clustered index scan in the case of an index-based table) is faster than table scan of a heap table

Regards

I am currently working on migrating queries from DB2 to SQL Server 2005.

Please refer to the following DB2 CREATE INDEX Query:

create index XFK_ACTV_CTC on ACTIVITY (
CNTX_TYP_CD ASC
)
allow reverse scans;

In IBM-DB2 the syntax “ALLOW REVERSE SCANS” – Specifies that an index can support both forward and reverse scans; that is, scanning of the index in the order that was defined at index creation time, and scanning in the opposite order.

I have been trying to search for an equivalent syntax in SQL Server 2005, however I am not able to do so.

Kindly help me in searching it.

Thanking you in anticipation.

Best Regards,
Amit

How does SQL Server 2005 issues a search inside a page that belongs to a clustered index? Does it scan the page or simply performs a binary search (I mean, since the records are ordered in a clustered index)?

Best regards!