How to count one column and adding up several columns at the same time using summarise in dplyr R

Question

I have this dataframe (there are 20 more columns like class but just for the example I show a short version):

 niv  ID  class1 class2 class3
  A   x    10      5      7
  A   y    15      3      1
  A   z    11      2      4
  B   u    17      7      3
  B   w    18      7      9

So I wanted to group_by(niv) and to summarise the ID column just counting it and class1, class2 and class3 columns to add up in just one command. I tried summarise_each but I can't sort out how to use it. Also the objetive would be to learn how to mix different funs in the summarise function when you have a dataframe with a lot of columns.

The summarise dataframe I need is this one:

 niv  n_ID  sum_class1 sum_class2 sum_class3
  A    3        36         10         12
  B    2        35         14         12

Thanks in advance.

It was mine, I sent it but it was incomplete so I erased it to post it with all the content. — Ernesto Riveros Barrientos, May 29 '20 at 04:38

akrun · Accepted Answer · 2020-05-29T04:43:53.537

1

We can use summarise with across

library(dplyr)
df %>%
   mutate(n_ID = 1) %>%
   group_by(niv) %>%
   summarise(across(starts_with('class'), sum))

With across, we can have multiple functions

iris %>%      
   summarise(across(where(is.factor), nlevels), 
            across(where(is.numeric), mean))
#  Species Sepal.Length Sepal.Width Petal.Length Petal.Width
#1       3     5.843333    3.057333        3.758    1.199333

data

df <- structure(list(niv = c("A", "A", "A", "B", "B"), ID = c("x", 
"y", "z", "u", "w"), class1 = c(10L, 15L, 11L, 17L, 18L), class2 = c(5L, 
3L, 2L, 7L, 7L), class3 = c(7L, 1L, 4L, 3L, 9L)), 
class = "data.frame", row.names = c(NA, -5L))

edited May 29 '20 at 04:43

answered May 28 '20 at 20:28

akrun

874,273
37
540
662

Thanks. I ran the code an I got this: `Error across (starts_with ("class"), sum): function "across" could not be found`. Also what I want for future exercises is to learn how to combine count and sum in one summarise function. – Ernesto Riveros Barrientos May 29 '20 at 04:36
@ErnestoRiverosBarrientos the function is in the new dplyr package version. If you need the bleeding edge, install the devel version – akrun May 29 '20 at 04:37
I understand. Can you help me where to find the devel version? – Ernesto Riveros Barrientos May 29 '20 at 04:44
1

@ErnestoRiverosBarrientos you can use `devtools::install_github("tidyverse/dplyr")` (from [here](https://github.com/tidyverse/dplyr) – akrun May 29 '20 at 04:45

score 0 · Answer 2 · answered May 28 '20 at 07:22

0

You can create n_ID column with all 1's and sum them with all 'class' columns.

library(dplyr)

df %>%
 mutate(n_ID = 1) %>%
 group_by(niv) %>%
 summarise_at(vars(n_ID, starts_with('class')), sum)

# A tibble: 2 x 5
#  niv    n_ID class1 class2 class3
#  <chr> <dbl>  <int>  <int>  <int>
#1 A         3     36     10     12
#2 B         2     35     14     12

data

df <- structure(list(niv = c("A", "A", "A", "B", "B"), ID = c("x", 
"y", "z", "u", "w"), class1 = c(10L, 15L, 11L, 17L, 18L), class2 = c(5L, 
3L, 2L, 7L, 7L), class3 = c(7L, 1L, 4L, 3L, 9L)), 
class = "data.frame", row.names = c(NA, -5L))

answered May 28 '20 at 07:22

Ronak Shah

377,200
20
156
213

That could solved it. But I also wanted to explore combining two funs in the summarise function. – Ernesto Riveros Barrientos May 28 '20 at 16:13
@ErnestoRiverosBarrientos I don't get it which 2 functions do you want to combine? You are applying only `sum` function to columns `class1` ,`class2`etc. – Ronak Shah May 29 '20 at 00:11
There are two functions I want to use: count niv and sum the other numeric columns. – Ernesto Riveros Barrientos May 29 '20 at 04:26
@ErnestoRiverosBarrientos Yes, so you are not actually combining two functions, right? Because you are counting only once. This gives you that I think. I am not exactly clear what you are looking for. – Ronak Shah May 29 '20 at 05:47
I think the `across` function suggested by @akrun can combine several functions in the summary function. – Ernesto Riveros Barrientos May 30 '20 at 05:18

How to count one column and adding up several columns at the same time using summarise in dplyr R

2 Answers2

data