I would like to calculate a rank-biserial correlation. But the (only it seems) package can\'t handle missing values that well. It has no built in "na.omit = TRUE" func