Say I'm doing some work which is closely related to someone's else's work. I want to study the same data that he studied. I know him well, and producing the data is no big deal but is time consuming, so I ask him to provide me with the data he had produced. If he agreed. Should I include him as a co-author?
The issue of whether or not merely providing data for a project constitutes "authorship" is a thorny one. Many medical journals, for example, are asserting that merely having generated data is not grounds for authorship all on its own.
So if he's just providing you with a data file, and that's the last you'll hear from him? In my mind, that lands you in the acknowledgements section of a paper.
On the other hand, if he's going back and forth retrieving data as you refine your question, providing advice on how it came about, what it means, etc. and may end up reviewing the draft of the paper, etc.? I'd make him an author. Essentially, I would extend him the chance to take up the slightly more burdensome mantle of authorship, and let him turn it down.