rgiordan/zaminfluence

NAs in the APIP cause error in `validate_APIP`

Closed this issue · 2 comments

A helpful user reports: When the linear approximation cannot produce a particular signal and the APIP is NA, zaminfluence fails with the following error:

Error in validate_APIP(new_APIP(n = n_drop, prop = n_drop/num_obs, inds = inds_drop)) : all(apip$inds > 0) is not TRUE

This is due to mishandling NA in validate_APIP, since I check for NULL but not NA.

Fixed in #26. I'll leave the issue open until there is a unit test.