1

Out of curiosity: I thought it was possible to use sum() to create a new variable in an R dataframe- the objective being calculating an overall score out of several single values. However sum() apparently sums all the values in a column and not just the values of a single case. What is the mechanism behind this and is there a function that adds the values as the simple addition does?

Daten <- data.frame(
  cases = c("first", "second", "third"), 
  values1= c(1,2,3),
  values2= c(27,19,34),
  values3= c(2,8,7)
)




Daten$valcomb = sum(Daten$values1,Daten$values2,Daten$values3)

Daten$valcomb2 = Daten$values1+Daten$values2+Daten$values3

print(Daten)

Output

   cases values1 values2 values3 valcomb valcomb2
1  first       1      27       2     103       30
2 second       2      19       8     103       29
3  third       3      34       7     103       44
Improve this question
1
  • Have you studied help("sum")? Commented Nov 11, 2021 at 12:36

2 Answers 2

Reset to default
2

If you are summing rows, you need to userowSums rather than sum. Obviously, you can't include the non-numeric cases column, so you need rowSums(Daten[-1]) to get the sums across the numeric rows.

within(Daten, sums <- rowSums(Daten[-1]))
#>    cases values1 values2 values3 sums
#> 1  first       1      27       2   30
#> 2 second       2      19       8   29
#> 3  third       3      34       7   44

Or, if you are using dplyr:

Daten %>% mutate(sums = rowSums(.[-1]))
#>    cases values1 values2 values3 sums
#> 1  first       1      27       2   30
#> 2 second       2      19       8   29
#> 3  third       3      34       7   44
Improve this answer
Sign up to request clarification or add additional context in comments.

1 Comment

Thanks for the reply. In the original data set I have a lot more columns though, so excluding with c(-1,-5, -6,...) works but wouldn't be very practical- how would you actively select the columns you want to sum?
1

This has nothing to do with data.frames but about the behaviour of the + as operator versus the behaviour of sum() as a function.

Using the + operator on vectors it will sum each vector element

c(1,2,3) + c(27,19,34) + c(2,8,7)
# [1] 30 29 44

Be aware though as it recycles the length if unequal in length where possible or only throwing a warning.

c(1,2,3,4) + c(27,19,34) + c(2,8,7)
# [1] 30 29 44 33
# Warning messages:
# 1: In c(1, 2, 3, 4) + c(27, 19, 34) :
#   longer object length is not a multiple of shorter object length
# 2: In c(1, 2, 3, 4) + c(27, 19, 34) + c(2, 8, 7) :
#   longer object length is not a multiple of shorter object length

While using the sum() function on vectors it will sum all values by definition of the function.

sum(c(1,2,3), c(27,19,34), c(2,8,7))
# [1] 103

Update

Besides some theory on + operator and sum function behaviour, here an example to use some functions to work with data.frame data.

library(dplyr)

Daten %>%
  # sums all columns that yield numeric values
  mutate(sum_all = rowSums(across(where(is.numeric)))) %>%
  # sums values1 and values3
  mutate(sum_1_3 = rowSums(across(c("values1", "values3"))))

#    cases values1 values2 values3 sum_all sum_1_3
# 1  first       1      27       2      30       3
# 2 second       2      19       8      29      10
# 3  third       3      34       7      44      10
Improve this answer

3 Comments

Thank you. That explains the logic of it very well and is exactly the kind of answer I was looking for. Is there a function that adds vectors the same way the + operator does? It doesn't matter in my example (or in my actual code) as there are only a few variables to add up, but I imagine if you'd want to add up a big number of vectors (vertically) it would be useful to have a function you can pass (values1 to values100) to instead of writing it all down variable by variable.
When you have your data in a data.frame and you want to either sum rows (of particular columns) or you want to sum particular columns, you better avoid doing those calculations on column subsets. Read into colSums and rowSums instead.
I added a small example in my answer.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.