Skip to contents

Taken from codebook You can use this function if some of your items have lost their attributes during wrangling Variables have to have the same name (Duh) and no attributes should be overwritten. But use with care. Similar to labelled::copy_labels().

Usage

rescue_attributes(df_no_attributes, df_with_attributes)

Arguments

df_no_attributes

the data frame with missing attributes

df_with_attributes

the data frame from which you want to restore attributes