I was thinking along the lines of:

SELECT *

FROM Employee E

    JOIN (

        SELECT ROW_NUMBER() OVER(PARTITION BY B.EmployeeId ORDER BY B.BillId DESC) AS Row

            ,B.BillId

            ,B.BillAmount

            ,B.BillMonth

            ,B.EmployeeId

        FROM EmployeeBill AS B

    ) D ON E.EmployeeId = D.EmployeeId

WHERE D.Row...

Try something like the following. The UPDLOCKs are important to avoid deadlocks.

UPDATE I

SET Status= 1

    ,@IdentPK = I.Identpk

FROM pIdent I

    JOIN (

        SELECT MAX(D1.IdentPK) AS IdentPK -- will ignore NULLs

        FROM (

            SELECT MIN(I1.IdentPK) AS...

I find your question confusing.

You are more likely to get appropriate help if sample data and expected results are provided.

Maybe you want something like:

DECLARE @EndDate smalldatetime

    ,@StartDate smalldatetime

SET @EndDate = CAST(CONVERT(varchar(6),...

Try looking up ROW_NUMBER() OVER(PARTITION BY ... in BOL.

My solution does need some error handling. (sp_releaseapplock in CATCH block etc)

Also, it would probably be better to use the SP name plus the @HashedBasketId for the resource just in...

If this is part of the

http://www.sqlservercentral.com/forums/shwmessage.aspx?forumid=360&messageid=346291

thread:

1. I would put the index on the FK and let the optimizer work out whether to use a MERGE or LOOP join based...

Opps,

See you have already done that.

Consistent use of application locks helps in this sort of situation.

Try something like:

SET QUOTED_IDENTIFIER ON

GO

SET ANSI_NULLS ON

GO

CREATE PROCEDURE [dbo].[ST_BASKET_INSERT]

    @BasketId int OUTPUT

    ,@HashedBasketId varchar

I suspect the problem is that CREATE LOGIN does not accept a variable for the login_name so a variable will not work within sp_executesql. (I do not have access to SQL...

A BatchID could also be generated from a dummy IDENTITY table. Something like.

CREATE TABLE dbo.BatchIDs

(

    BatchId int IDENTITY NOT NULL

)

GO

DECLARE @BatchID int

BEGIN TRAN

INSERT INTO dbo.BatchIDs DEFAULT VALUES

ROLLBACK

SELECT @BatchID = SCOPE_IDENTITY()

-- SELECT...

Get rid of Enumerations and have separate tables for States, Countries etc.

Maybe:

SELECT @result = DATEDIFF(day, MIN(StartDATE), MAX(EndDate))

FROM @MyTable

SELECT REPLACE(STR(MONTH(GETDATE()), 2), ' ', '0')

In this situation, Remi's outline solution is best.

You should however use an UPDLOCK on the select in the derived table. Ensuring that only one of the processes can read the...

I think you need to put an UPDLOCK in the derived table in order to guarantee that only one transaction can read the rows. Also, there may as well be...