First misconception is that white plays one opening while the other side is playing another. This is not correct. The opening is defined by both black and white moves. Players don't independently develop their moves. Look at Sicilian which can illustrate this. The Ruy Lopez doesn't have c6 which is found in the Caro Kann. White can play c3 though. f6 would be an inaccuracy and bad development for black early on.
What I would do is look up games from a database. You will see some moves are more popular than the engine top choice. I would start there and study those. I wouldn't limit your games to only games you have played in the past. It will be highly inadequate.
I am not a player who is good enough to switch between openings. I need to play what I know. When preparing for a match and I know my opponents favourite opening is there any engine which an engine can match what the their prefererences to mine. For example I play the ruy Lopez and I know my opponent likes the caro kann. I want an engine which could look at how both players might go about getting the set ups they want which might not always follow the optimal line. In the real world players dont always choose the best move, but they try to get familiar positions. Both players want to reach the set ups they are familiar with. I want to see how these positions might best be reached not always by the optimal move but where both players are trying to reach particular set ups.
Can an engine do this. If not does anybody have suggestions how such games might be prepared for.
many thanks
Edwin