If you’re a GM (or top level CC player) opening books are pretty much worthless because any opinion on a variation more than a month or two old is outdated. So what you need is Chessbase (not the light version) and two strong engines; most likely the latest version of Fritz and Rybka. The reason for two is because sometimes Rybka is overly optimistic and Fritz too cautious. Of course your database must be kept up to date by downloading games from The Week in Chess, opening surveys from Chessbase Magazine and updates from Chess Publishing.
Then you are prepared to set up engine tournaments to be played overnight from the opening positions that you are interested in. That way you will be able to see how best to approach play in those positions. Unfortunately, unless you are a GM, no database, engine evaluations or game results from engine tournaments are going to help you correctly evaluate all the nuances of two or three unclear or equally evaluated positions.
GM Vladimir Bagirov opined that after three successive defeats with any opening, then no matter how much you like it, it’s time to replace it. Also, if you are a GM, you have to be aware of the fact that most of your peers will only use an opening variation for a couple of games. What they do is choose a complicated variation, check it with an engine then play it 2-3 times and then abandon it because somebody has probably found a refutation. Most GM's maintain a solid mainline repertoire in which they are constantly looking for novelties. But that’s GM’s. What about the rest of us?
Fortunately all we need is one solid line against most openings/defenses. For example, If you are planning to meet 1.e4 with ...e5, all you need is one solid line against the Ruy Lopez. Likewise one solid variation against the Giuoco Piano and King’s Gambit will suffice. You would follow the same procedure for White. You can determine this information by looking at your db statistics.
Once you’ve established your main lines you can do a search for games using those lines, print out 20-25 games and play over them paying attention to the general themes, patterns and any tactical motifs that arise. Just make sure you play over the entire game to get a feel for the game as a whole. You may also want a book on your particular opening/defense or main variation so you can get a good verbal description of the main ideas. That way you will know what you are trying to accomplish and it will help you better understand the ideas behind the moves played in your sample games.
Many lower rated players feel they need to avoid openings like the Ruy Lopez and Nimzo-Indian because they think those openings are too complicated for them to play well. That is not true. It’s going to take no more study to learn to play those lines properly than it is to play some second rate opening/defense where one slip is likely to cost you the game. It’s just not possible to win very often when you are constantly getting positions where you have to find the best move nearly every turn.