34

with str(data) I get the headof the levels (1-2 values)

fac1: Factor w/ 2  levels ... :
fac2: Factor w/ 5  levels ... :
fac3: Factor w/ 20 levels ... :
val: num ...

with dplyr::glimpse(data) I get more values, but no infos about number/values of factor-levels. Is there an automatic way to get all level informations of all factor vars in a data.frame? A short form with more info for

levels(data$fac1)
levels(data$fac2)
levels(data$fac3)

or more precisely a elegant version for something like

for (n in names(data))
  if (is.factor(data[[n]])) {
    print(n)
    print(levels(data[[n]]))
  }

thx Christof

ckluss
  • 1,477
  • 4
  • 21
  • 33

8 Answers8

33

Here are some options. We loop through the 'data' with sapply and get the levels of each column (assuming that all the columns are factor class)

sapply(data, levels)

Or if we need to pipe (%>%) it, this can be done as

library(dplyr)
data %>% 
     sapply(levels)

Or another option is summarise_each from dplyr where we specify the levels within the funs.

 data %>%
      summarise_each(funs(list(levels(.))))
akrun
  • 874,273
  • 37
  • 540
  • 662
9

If your problem is specifically to output a list of all levels for a factor, then I have found a simple solution using :

unique(df$x)

For instance, for the infamous iris dataset:

unique(iris$Species)

Djamil Lakhdar-Hamina
  • 111
  • 1
  • 2
6

Or using purrr:

data %>% purrr::map(levels)

Or to first factorize everything:

data %>% dplyr::mutate_all(as.factor) %>% purrr::map(levels)

And answering the question about how to get the lengths:

data %>% map(levels) %>% map(length)
Paul
  • 3,920
  • 31
  • 29
Amit Kohli
  • 2,860
  • 2
  • 24
  • 44
4

A simpler method is to use the sqldf package and use a select distinct statement. This makes it easier to automatically get the names of factor levels and then specify as levels to other columns/variables.

Generic code snippet is:

library(sqldf)
    array_name = sqldf("select DISTINCT *colname1* as '*column_title*' from *table_name*")

Sample code using iris dataset:

df1 = iris
factor1 <- sqldf("select distinct Species as 'flower_type' from df1")
factor1    ## to print the names of factors

Output:

  flower_type
1      setosa
2  versicolor
3   virginica
doncherry
  • 259
  • 3
  • 14
Ann Rajaram
  • 61
  • 2
2

In case you want to display factor levels only for thos columns which are declared as.factor, you can use:

lapply(df[sapply(df, is.factor)], levels)
Peter
  • 2,120
  • 2
  • 19
  • 33
0

Alternate option to get length of levels in a 'data'.frame:

data_levels_length <- sapply(seq(1, ncol(data)), function(x){
  length(levels(data[,x]))
})
Jay J
  • 1
  • 1
0

As a long data frame (tibble):

df %>% gather(name, value) %>% count(name, value)

This converts all the columns into name-value pairs, and then counts the unique levels.

Subset column types with something like:

df %>% select_if(is.character) %>% ...

Via https://stackoverflow.com/a/47122651/3217870

Paul
  • 3,920
  • 31
  • 29
0
library(dplyr) #for all the following

df$factor %>% unique() %>% str()

lists and counts the frequency of levels of a specific variable

count(df,variable)

returns a table with the levels of a specific variable and its frequency. the number of rows will inform how many levels there are for this variable.

count(df,across())

returns a table of all variables levels that co-occur in observations and the frequency of all different combinations

Driele Delanira
  • 1
  • 2