In a recent blog, Kalen Delaney said she is frequently asked why is there a need to understand the internals of SQL Server. I think this question illustrates why. The key to getting the right answer is an understanding of the sequence of events during the query processing - the HAVING clause is processed after the GROUP BY and acts on the result set from the earlier phases.

I learned this from Itzik Ben-Gan's book 'Inside Microsoft SQL Server 2008: T-SQL Querying'. Chapter 1, Logical Query Processing, has a great section on the sequence of events that occurs when the query processor gets to work. I really recommend this book to anyone who is struggling to understand the explanation for this QotD.