Cummulative difference In R-CodePudding

What is the opposite of cumsum() in R

a = c(2, 5, 8)
cumsum = c(2, 7, 15)
cumdiff = c(2, 3, 1)

Because 5-2 = 3 and 8-7 = 1.

Is there a package that can be used for this one in R?

CodePudding user response：

In base R:

cumdiff <- function(x) x - head(c(0, cumsum(x)), -1)
#cumdiff <- function(x) x - c(0, cumsum(x))[-(length(x)   1)]

In dplyr, with lag and default = 0:

library(dplyr)
cumdiff <- function(x) x - lag(cumsum(x), default = 0)

Result:

a = c(2,5,8)
cumdiff(a)
#[1] 2 3 1

CodePudding user response：

If by cumdiff you mean the current value in the vector minus the cumsum of the previous values, you could use the below:

a = c(2,5,8)
cumdiff <- function(x){
  out <- x - lag(cumsum(x))
  out[1] <- x[1]
  return(out)
}
cumdiff(a)
[1] 2 3 1


library(dplyr)
tibble(a) %>% 
  mutate(cumsum = cumsum(a),
         cumdiff = cumdiff(a))
#> # A tibble: 3 × 3
#>       a cumsum cumdiff
#>   <dbl>  <dbl>   <dbl>
#> 1     2      2       2
#> 2     5      7       3
#> 3     8     15       1

^{Created on 2022-09-08 by the reprex package (v2.0.1)}

CodePudding user response：

library(dplyr)
data.frame(a)  %>% 
  mutate(
         cumdiff = a - lag(cumsum(a)),
         cumdiff = ifelse(is.na(cumdiff), a, cumdiff)
         )
  a cumdiff
1 2       2
2 5       3
3 8       1