R How to take Median of Rows in Dataframe

Question

I am wondering if there is any way to a median of the rows in a data frame. I understand the function rowmeans exists, but I do not believe there is a row median function. I would like to store the results in a new column in the dataframe. Here is my example

I tried to look online. There was one mention of row medians, but I could not find the function in R.

 C1<-c(3,2,4,4,5)
   C2<-c(3,7,3,4,5)
   C3<-c(5,4,3,6,3)
   DF <- data.frame(ID=c("A","B","C","D","E"),C1=C1,C2=C2,C3=C3)

   DF 


  # This is as far as I have gotten, but not streamlined

  MA <- median(C(3, 3, 5). na.rm = T)   # A
  MB <- median(C(2, 7, 4). na.rm = T)   # B
  MC <- median(C(4, 3, 3). na.rm = T)   # C
  MD <- median(C(4, 4, 6). na.rm = T)   # 4
  ME <- median(C(5, 5, 3). na.rm = T)   # E

  CM <- c(MA, MB, MC, MD, ME)C1<-c(3,2,4,4,5)


   ID C1 C2 C3
  1  A  3  3  5
  2  B  2  7  4
  3  C  4  3  3
  4  D  4  4  6
  5  E  5  5  3

   ID C1 C2 C3  CM
  1  A  3  3  5
  2  B  2  7  4
  3  C  4  3  3
  4  D  4  4  6
  5  E  5  5  3

Is there anyway I can streamline the process so it would be like DF$CM <- median(...

There's function `matrixStats::rowMedians`. [Related: how to add median value to rows?](https://stackoverflow.com/questions/28077887/how-to-add-median-value-to-rows); [How to calculate row medians efficiently with data.table](https://stackoverflow.com/questions/48885416/how-to-calculate-row-medians-efficiently-with-data-table) — pogibas, Jan 25 '19 at 13:50
Possible duplicate of [Find median of every row using matrixStats::rowMedians](https://stackoverflow.com/questions/51525152/find-median-of-every-row-using-matrixstatsrowmedians) — pogibas, Jan 25 '19 at 13:52
As noted by jogo, your go to function for anything rowwise should be `apply` with the dimension set to 1. — iod, Jan 25 '19 at 14:38

score 16 · Answer 1 · edited May 10 '21 at 09:24

16

To calculate the median of df, you can do the following

df$median = apply(df, 1, median, na.rm=T)

edited May 10 '21 at 09:24

Mathias711

6,568
4
41
58

answered Mar 12 '20 at 20:35

Jiaqi

486
4
5

score 6 · Answer 2 · answered Jan 25 '19 at 15:36

If you would like to use dplyr, you can find an example here, especially mpalanco's answer. Briefly, after using rowwise to indicate that the operation should be applied by row (rather than to the entire data frame, as by default), you can use mutate to calculate and name a new column off of a selection of existing columns. Check out the documentation on each of those functions for more details.

E.g.,

library(dplyr)

DF %>% 
  rowwise() %>% 
  mutate(CM = median(c(C1, C2, C3), na.rm = TRUE))

will yield the output:

# A tibble: 5 x 5
  ID       C1    C2    C3    CM
  <fct> <dbl> <dbl> <dbl> <dbl>
1 A         3     3     5     3
2 B         2     7     4     4
3 C         4     3     3     3
4 D         4     4     6     4
5 E         5     5     3     5

row-wise is painfully slow with large datasets – Markm0705 Oct 22 '21 at 01:05 — Markm0705, Oct 22 '21 at 01:05

score 5 · Answer 3 · answered May 10 '21 at 09:52

Just a little bit more flexible and up to date. We use c_across with rowwise function and it allows to use tidy-select semantics. Here we choose where to specify we only want the numeric column to calculate the median.

library(dplyr)

DF %>%
  rowwise() %>%
  mutate(med = median(c_across(where(is.numeric)), na.rm = TRUE))

# A tibble: 5 x 5
# Rowwise: 
  ID       C1    C2    C3   med
  <chr> <dbl> <dbl> <dbl> <dbl>
1 A         3     3     5     3
2 B         2     7     4     4
3 C         4     3     3     3
4 D         4     4     6     4
5 E         5     5     3     5

JAdel · Answer 4 · 2022-12-16T16:35:26.260

0

Oneliner that allows you to pick your desired columns:

apply(DF[, c("C1", "C2", "C3")], 1, median)

  ID C1 C2 C3 CM
1  A  3  3  5  3
2  B  2  7  4  4
3  C  4  3  3  3
4  D  4  4  6  4
5  E  5  5  3  5

edited Dec 16 '22 at 16:35

answered Dec 16 '22 at 16:13

JAdel

1,309
1
7
24

R How to take Median of Rows in Dataframe

4 Answers4