More DBA job interview questions and answers at
http://dba.fyicenter.com/Interview-Questions/
(Continued from previous question...)
Which of the queries is the best one for Janice to use to return the information she has been requested to provide?
Janice has two tables, Employees and Orders. She has been asked to provide a report of the number of orders processed by each employee
in the month of June. If an employee didn’t process any orders, the report should reflect the employee’s name and a zero for the number of orders. Which of the queries is the best one for Janice to use to return the information she has been requested to provide?
Answer
SELECT
E.LastName + ', ' + E.FirstName AS [Employee Name]
, ISNULL(O.[# of Orders], 0) [# of Orders]
FROM dbo.Employees E
LEFT JOIN (SELECT
EmployeeID
, COUNT(*) [# of Orders]
FROM dbo.Orders
WHERE OrderDate >= '20060601'
AND OrderDate < '20060701'
GROUP BY EmployeeID) O
ON E.EmployeeID = O.EmployeeID
ORDER BY [Employee Name]
Explanation
While it would seem BETWEEN would save Janice a bit of typing, there is a problem with it. BETWEEN corresponds, based on these queries to the same as:
WHERE OrderDate >= ‘20060601’ AND OrderDat e<= ‘20060701’
It’s the latter one that causes the query to be incorrect as it would potentially include orders placed on July 1st at midnight (20060701 corresponds to July 1, 2006 at 00:00:00). While this would be unlikely in a small retail environment, it is entirely possible in larger operations, especially those which are international in scope. Therefore, the use of BETWEEN in this case is not appropriate.
With respect to the GROUP BY, it must appear in the subquery on the Orders table, as given in the answer. When it occurs after the LEFT JOIN, the rows where there is an employee but no order for the month will be lost. Since Janice must report on employees who had no sales, this is unacceptable.
(Continued on next question...)
