T-SQL programmers often need to handle data that comes in a denormalized form. Certain tables might appear in a denormalized form because they store data that was imported from legacy systems, for example, or from text files. Also, many online transaction processing (OLTP) environments aren't properly normalized even though they should be. Let's look at an example problem that involves temperature measurement samples stored in a denormalized form. I'd like to thank SQL Server MVPs Linda Wierzbicki and Steve Kass, who came up with some of the tricks that I use in this article.