R - Loop to subtract each successive value from first in a series - r

Question

I am tinkering around with this loop (im new to writing loops but trying to learn).
I am aiming when x == 1, on first 1 match, store the value of z, then on each successive z value subtract that z value from the first value. If x == 0 then it will do nothing (not sure if i have to tell the code to do nothing when x ==0?)
This is my dummy data:
x <- c(0,0,1,1,1,0,1,1,1,1,0,0,0)
z <- c(10,34,56,43,56,98,78,98,23,21,45,65,78)
df <- data.frame(x,z)
for (i in 1:nrow(df)) {
if (df$x[i] == 1)
first_price <- df$z[i]
df$output <- first_price - df$z
}
}
I have my if (df$x == 1)
Then I want to save the first price... so first_price <- df$z[i]
The i in here, that means the first in the series right?
Then for my output... i wanted to subtract the first price from each successive price. If I fix the first price with [i] is this the correct way? And if I leave df$z would that then take the next price each time in the loop and subtract from
first_price <- df$z[i]?
Heres a visual:
******Progress****
> for (i in 1:nrow(df)) {
+ if (df$x[i] == 1) {
+ first_price <- df$z[1]
+ df$output <- first_price - df$z
+ }
+ }
> df$output
[1] 0 -24 -46 -33 -46 -88 -68 -88 -13 -17 -35 -55 -68
If i add [1] which is assigning the first element in df$z this actually fixes the first element and then subtracts each successive, now It needs to be rule based and understand that this is only to be the case when df$x == 1

Answer 1

This should work for you
library(dplyr)
library(data.table)
ans <- df %>%
mutate(originalrow = row_number()) %>% # original row position
group_by(rleid(x)) %>%
mutate(ans = first(z) - z) %>%
filter(x==1)
# # A tibble: 7 x 5
# # Groups: rleid(x) [2]
# x z originalrow `rleid(x)` ans
# <dbl> <dbl> <int> <int> <dbl>
# 1 1 56 3 2 0
# 2 1 43 4 2 13
# 3 1 56 5 2 0
# 4 1 78 7 4 0
# 5 1 98 8 4 -20
# 6 1 23 9 4 55
# 7 1 21 10 4 57
vans <- ans$ans
# [1] 0 13 0 0 -20 55 57
EDIT
To keep all rows, and outputting 0 where x==0
ans <- df %>%
mutate(originalrow = row_number()) %>%
group_by(rleid(x)) %>%
mutate(ans = ifelse(x==0, 0, first(z) - z))
# # A tibble: 13 x 5
# # Groups: rleid(x) [5]
# x z originalrow `rleid(x)` ans
# <dbl> <dbl> <int> <int> <dbl>
# 1 0 10 1 1 0
# 2 0 34 2 1 0
# 3 1 56 3 2 0
# 4 1 43 4 2 13
# 5 1 56 5 2 0
# 6 0 98 6 3 0
# 7 1 78 7 4 0
# 8 1 98 8 4 -20
# 9 1 23 9 4 55
# 10 1 21 10 4 57
# 11 0 45 11 5 0
# 12 0 65 12 5 0
# 13 0 78 13 5 0