This is a good solution.
However this is close to na.rm = TRUE in R, which skipmissing has taken pains to avoid.
I wonder if there is some room for a skipmissing function like
miss_x, miss_y = skipmissing(x, y) that has the missing elements for both dropped?