CODEWITHSUNDEEP
rdplyr
Bryan Adams
Bryan Adams

Reputation: 174

Count the number of columns in a row with a specific value

I have a data set with numerical responses to several questions. I would like to know the number of times a person answers a question with a value of 1,2...

Here is an example of the data:

df=data.frame("Person"=c("person a", "person b"),
              "Q1"=c(2,2),"Q2"=c(1,2),"Q3"=c(1,1))

Which looks like this:

 Person    Q1 Q2 Q3
 person a  2  1  1
 person b  2  2  1

I would like this and would prefer to use dplyr:

 Person    Q1 Q2 Q3  Total.1  Total.2
 person a  2  1  1   2        1
 person b  2  2  1   1        2

Upvotes: 1

Views: 137

Answers (3)

akrun
akrun

Reputation: 887881

Here is one option with tidyverse

library(tidyverse)
df %>% 
   mutate(Total = pmap(.[-1],  ~ 
              c(...) %>% 
                paste0("Total.", .) %>%
                table %>%
                as.list %>% 
                as_tibble )) %>%
#       unnest
#    Person Q1 Q2 Q3 Total.1 Total.2
#1 person a  2  1  1       2       1
#2 person b  2  2  1       1       2

Or another way is

df %>%
   mutate(Total = pmap(.[-1], ~
                    c(...) %>% 
                     table %>%
                     toString)) %>% 
   separate(Total, into = c("Total.1", "Total.2"))
#    Person Q1 Q2 Q3 Total.1 Total.2
#1 person a  2  1  1       2       1
#2 person b  2  2  1       1       2

Upvotes: 0

Ronak Shah
Ronak Shah

Reputation: 389265

The base R approach suggested by @dww is quite simple and straight forward. However, if you prefer dplyr approach we can use rowwise and do to calculate occurrence of 1 and 2 respectively. 

library(dplyr)

df %>%
  rowwise() %>%
  do( (.) %>% as.data.frame %>% 
             mutate(Total.1 = sum(.==1), 
                    Total.2 = sum(.==2)))

# Person      Q1    Q2    Q3 Total.1 Total.2
#  <fct>    <dbl> <dbl> <dbl>   <int>   <int>
#1 person a     2     1     1       2       1
#2 person b     2     2     1       1       2

A base R approach using apply

df[c("Total.1", "Total.2")] <- t(apply(df, 1, function(x) c(sum(x==1), sum(x==2))))

df
#    Person Q1 Q2 Q3 Total.1 Total.2
#1 person a  2  1  1       2       1
#2 person b  2  2  1       1       2

Upvotes: 2

dww
dww

Reputation: 31454

No need for dplyr. In base R it is quite simple

df = cbind(df, Total.1 = rowSums(df[,-1]==1), Total.2 = rowSums(df[,-1]==2))

Upvotes: 2

Related Questions