Is there a way in R for doing a pairwise-weighted covariance matrix?

Question

I have a survey with a lot of numeric variables (both continuous and dummy-binary) and more than 800 observations. Of course, there is missing data for most of the variables