SELECT with Joins returning ambiquous data
Posted on 2006-06-28
I have the following query (stripped down somewhat-it is part of a UNION)
MAX(tblPayrollRegisterPayments.PayTypeCode) AS PayTypeCode,
SUM(tblPayrollRegisterPayments.Hours) AS OTHours,
SUM(tblPayrollRegisterPayments.Amount) AS OTPay,
0 AS HourlyPayRate,
0 AS YTDCompHours,
0 AS YTDHCompHours
FROM tblPayrollRegisterPayments INNER JOIN
HumanResourcesTest.dbo.Employee E ON tblPayrollRegisterPayments.EmpID = E.emp_no INNER JOIN
tblPayrollRegisterEmployee ON tblPayrollRegisterPayments.EmpID = tblPayrollRegisterEmployee.EmpID AND
tblPayrollRegisterPayments.PayPeriodEndingDate = tblPayrollRegisterEmployee.PayPeriodEndingDate
WHERE tblPayrollRegisterEmployee.DepartmentCode LIKE CASE @Facility
WHEN 1 THEN '494%'
WHEN 2 THEN '555%'
AND (tblPayrollRegisterPayments.PayTypeCode = 'Overtime')
AND (tblPayrollRegisterPayments.PayPeriodEndingDate BETWEEN @StartDate AND @EndDate)
GROUP BY E.Classification, E.Lname, E.Fname, E.emp_no, tblPayrollRegisterEmployee.DepartmentCode
tblPayrollRegisterEmployee is the 'Parent' table and tblPayrollRegisterPayments is its child.
tblPayrollRegisterEmployee contains and employee ID which I can use to get additional information from Employee (name mainly)
The problem: The Employee table I am using is a 'temporary' one I built for the purpose of this report. It contains merged data from two physical locations plus 'archived' data. The archived data is needed because the report produced from this data is grouped by classification and classification is in the employee table and if archived (terminated) records were not included, their data would not be part of the output. Rather than tackling the problem from the perspective of reloading all the employee data, I would rather work it out in the procedure.
Anyway, I digressed. The problem is that we have a number of employees who terminated employment then rehired. Also, because of dual functions, some employees are in the employee database at both locations. When that is the case, I double (or even triple) the totals. Can someone suggest an easy way (maybe changing the JOIN statements?) to insure that for each payment, I only retrieve a single employee record.
A couple of 'rules' that might be able to be used: If they were terminated and rehired, there is a HireDate in the employee table and I could possibly get the most recent using MAX(E.HireDate) in some way. If they are in both loactions' tables, there is a possibility (not guarantee) that E.Department (which I am not currently using) would be different and the 'real' employee record's e.department would be the same as tblPayrollRegisterEmployee.DepartmentCode
I am not too concerned really in getting the 'correct' employee record, since what I am mainly interested in is the e.Classification and if they terminated and were rehired, there is a 99% probability that they hired back in the same classification and there is a 99% probability that the classification between locations is the same.