As I discussed in my last two columns, one way that the query optimizer in SQL Server 2000 and 7.0 is so much more complex and powerful than in previous releases is that it has many more query-processing options to choose from. In particular, besides a nested-loop join, which was the only join-processing technique originally available, SQL Server 7.0 introduced merge joins and hash joins. This month, I continue my examination of query plans available to the optimizer by telling you about situations in which each join type is useful.