CODEWITHSUNDEEP
rdplyr
Ignacio
Ignacio

Reputation: 7928

drop columns that take less than n values?

Suppose i have a data frame like the following:

df <- data.frame(v1 = sample(1:10, 100, replace = T), v2 = sample(LETTERS, 100, replace = T),
                 V3 = sample(letters, 100, replace = T), v4 = sample(1:15, 100, replace = T))

I would like to create a new data frame df2 only includes the columns that take more than 10 values. So, in this example it would be v2, v3, and v4. How can I do that? In practice my data frame has thousands of columns.

I tried this:

df2 <- df %>% select(which(length(unique(.))>10))

Upvotes: 2

Views: 611

Answers (2)

akuiper
akuiper

Reputation: 214937

Alternatively, you can use select_if() from dplyr where you can pass a function as predicate to select columns:

library(dplyr)
df %>% select_if(function(col) n_distinct(col) > 10)

#    v2 V3 v4
#1    T  a 12
#2    R  k  7
#3    L  l  1
# ...

Or using select with where in dplyr version >=1.00

df  %>%
     select(where(~ n_distinct(.) > 10))

Upvotes: 8

akaDrHouse
akaDrHouse

Reputation: 2240

Clunky but it works...

x<-as.data.frame(t(apply(df,2,function(x) length(x[unique(x)]))>10))

df[,names(x[,x>0])]

Upvotes: 0

Related Questions