Closed reiniervlinschoten closed 1 year ago
Hi! A quick fix could be to convert the factor to a numeric variable before imputation, and recoding back afterwards?
Thanks.
I patched mice.impute.2lonly.pmm()
in mice 3.15.4
, so we should not see the warning anymore.
A weird bug is occuring, maybe related to #410. When imputing a factor variable that should be constant within a class (e.g. smoking for a patient with longitudinal measurements), something goes wrong.
Reprex (from the vignette of the 2lonly.pmm function) which gives as a warning: Warning message:
And the resulting dataframe has missing values in column V.