Maybe not books per se, but I often find more exhaustive treatments and variations on a proof in survey papers.