In 2008, Tony Reese presciently told us that the case law on fair use "transformativeness" showed a trend towards favoring transformative purpose over transforming content, so that exact reproduction could have a very good shot at fair use. Today's example, via Eric Goldman, involves Westlaw and Lexis's creation of databases containing the full text (minus some privacy redactions) of legal filings. In a quick opinion, the court deems this conduct clearly transformative--it results in the creation of a completely new thing, a legal research database--and not harmful despite being commercial and involving full copies (or as near as makes no difference). At this point, I think it's safe to say that it's easier to win a fair use case by engaging in large-scale, wholesale copying to create a database than it is to win a fair use case by altering the content of a single work--there are plenty of cases in the latter category, of course, but there are also cases finding infringement, and so far there aren't any in the former category. I would be surprised if the Authors' Guild v. Google case became the first.
Critics of so-called "expansive" fair use holdings sometimes argue that (1) favoring databases is an unjustified extrapolation from Campbell, which after all was a content-transformativeness case, and (2) that conceptually, "transformativeness" overlaps with the derivative works right, which also speaks of how a work may be "transformed" or adapted. (1) and (2) are, it seems to me, really in conflict--any need to constrain content-transformativeness for fair use purposes in order to protect the derivative works right doesn't apply to purpose-transformativeness, which generally needs to work with reproductions to achieve its aims. Anyway, Reese's careful framework seems to me to explain the subsequent six years of cases too.