Wipe out and re-calculate attacks and move in each node - can I do this smarter?

[evert823]

In each node I calculate attacks + candidate moves.
So when I have a candidate move in the current position, I create the position resulting from doing that move, and then I re-calculate attacks + candidate moves.

For example this position:
[d]k6r/br3p1p/6p1/4p3/P2p4/1Q1P4/2PK2PP/7B w - - 0 1

Now when we look at candidate moves 1. h3 and 1. g3, we see a vast difference in impact on the calculated information.
After 1. h3 almost everything remains the same. After 1. g3 we have a new pin on the Rook and the Queen is in considerably less danger.

In my program I have a 100% safe approach: I wipe out all information and re-calculate everything. So I have no bugs but I am doing more work.

Question:
Is it common to try to re-use calculated information (attacks + candidate moves) or re-calculate it only partially?