Error integer overflow - use sum(as.numeric(.))[1] NA in gapminder dataset

Question

I was playing with the gapminder dataset, and when trying to sum the population, an integer variable, an error

library(gapminder)

gapminder %>%
        group_by(year) %>% 
        summarise(totalPop = sum(pop))

Shows NA as results

Then, if try

mean(gapminder$pop)
[1] 29601212 # ok, as expected

but if

sum(gapminder$pop)
integer overflow - use sum(as.numeric(.))[1] NA

I read the What is integer overflow in R and how can it happen?

so, I tried

sum(gapminder$pop / 1)
[1] 50440465801

Any explanation why sum doesn't work, but sum / 1 did?

In brief: why?

gapminder %>%
    group_by(year) %>% 
    summarise(totalPop = sum(pop))

but

gapminder %>%
    group_by(year) %>% 
    summarise(totalPop = sum(pop/1))

score 2 · Accepted Answer · answered Dec 10 '17 at 00:15

2

Dividing by 1 coerces it numeric.

library(gapminder)
class(sum(gapminder$pop))
[1] "integer"
Warning message:
In sum(gapminder$pop) : integer overflow - use sum(as.numeric(.))

class(sum(gapminder$pop/1))
[1] "numeric"

answered Dec 10 '17 at 00:15

jmuhlenkamp

2,102
1
14
37

thanks, makes sense. Anyway, any clue why sum(pop) gives NA? – sergiouribe Dec 10 '17 at 01:19
sum(pop) gives NA because of the integer overflow. I.e. the sum of pop is larger than range of allowable integers. Numeric is not restricted like the range of integers. – jmuhlenkamp Dec 10 '17 at 03:51
thanks. Learned that .Machine$integer.max #> [1] 2147483647 – sergiouribe Dec 11 '17 at 18:08
Is there a more formal way to do this? Dividing by 1 seems like a work-around. – BenW Dec 13 '20 at 01:06

Error integer overflow - use sum(as.numeric(.))[1] NA in gapminder dataset

1 Answers1