Combine Rarely Used Factor Levels Across all Factors in Data Frame

问题

I would like to combine factor levels that are used fewer than 5 times for each factor in a dataset containing many different factors. While I understand that the fct_lump() function in the forcats package can help me achieve this for a single factor, is there a function where I can apply the fct_lump() function over all the factors in my dataset?

回答1:

We can check whether the column is factor with mutate_if and apply fct_lump

library(dplyr)
library(forcats)
df1 %>%
     mutate_if(is.factor, fct_lump)

---

Or in base R

i1 <- sapply(df1, is.factor)
df1[i1] <- lapply(df1[i1], fct_lump)

来源：https://stackoverflow.com/questions/60956344/combine-rarely-used-factor-levels-across-all-factors-in-data-frame