Creating a variable that depends on it's previous value as well as other variables

[weetbix_warrior]

I want to create a variable that depends on itself as well as other variable values.
I think the simplest that I can describe the problem is as per below.

Variable want is a cumulative sum. If the variable c == F then it adds b otherwise it adds a

• so the first 4 elements is the regular cumsum(1:4)
• but then the next element is want[4] + a[5] which is 11
• next element is want[5] + b[6] which is 17
• and so on

df <- data.frame(
  a = rep(1,10),
  b = 1:10,
  c = c(F,F,F,F,T,F,T,T,F,T),
  want = c(1,3,6,10,11,17,18,19,28,29)
)

[FJCC]

library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
df <- data.frame(
  a = rep(1,10),
  b = 1:10,
  c = c(F,F,F,F,T,F,T,T,F,T),
  want = c(1,3,6,10,11,17,18,19,28,29)
)
df
#>    a  b     c want
#> 1  1  1 FALSE    1
#> 2  1  2 FALSE    3
#> 3  1  3 FALSE    6
#> 4  1  4 FALSE   10
#> 5  1  5  TRUE   11
#> 6  1  6 FALSE   17
#> 7  1  7  TRUE   18
#> 8  1  8  TRUE   19
#> 9  1  9 FALSE   28
#> 10 1 10  TRUE   29
df |> mutate(want2 = cumsum(ifelse(c,a,b)))
#>    a  b     c want want2
#> 1  1  1 FALSE    1     1
#> 2  1  2 FALSE    3     3
#> 3  1  3 FALSE    6     6
#> 4  1  4 FALSE   10    10
#> 5  1  5  TRUE   11    11
#> 6  1  6 FALSE   17    17
#> 7  1  7  TRUE   18    18
#> 8  1  8  TRUE   19    19
#> 9  1  9 FALSE   28    28
#> 10 1 10  TRUE   29    29

^{Created on 2024-02-28 with reprex v2.0.2}

[weetbix_warrior]

Thanks for this reply @FJCC and for making me look silly .

I simplified the problem too much so can I please ask you to look at another small extension.

Variable want is a cumulative sum. If the variable c == F then it adds b otherwise it adds a

• so the first 3 elements is the regular cumsum(1:3)
• next because the previous value is > 5 set want[4] to 0
• next condition is T so want[5] <- want[4] + 1
• next element is want[5] + b[6] which is 7
• because the previous want value is greater than 5 set 'want[7]` to 0
• and so on

df <- data.frame(
  a = rep(1,10),
  b = 1:10,
  c = c(F,F,F,F,T,F,T,T,F,T),
  want = c(1,3,6,0,1,7,0,1,10,0)
)

[FJCC]

This is a more complicated solution. It relies on the fact that a FALSE comparison returns 0 and a TRUE comparison returns 1.

df <- data.frame(
  a = rep(1,10),
  b = 1:10,
  c = c(F,F,F,F,T,F,T,T,F,T),
  want = c(1,3,6,0,1,7,0,1,10,0)
)
df
#>    a  b     c want
#> 1  1  1 FALSE    1
#> 2  1  2 FALSE    3
#> 3  1  3 FALSE    6
#> 4  1  4 FALSE    0
#> 5  1  5  TRUE    1
#> 6  1  6 FALSE    7
#> 7  1  7  TRUE    0
#> 8  1  8  TRUE    1
#> 9  1  9 FALSE   10
#> 10 1 10  TRUE    0
df$want2 <- NA
for(i in 1:nrow(df)) {
  if(i == 1) df$want2[1] = df$a[1] * df$c[1] + df$b[1] * !df$c[1]
  else df$want2[i] = (df$want2[i-1] + (df$a[i] * df$c[i]) + (df$b[i] * !df$c[i])) * (df$want2[i-1] <= 5)
  }
df
#>    a  b     c want want2
#> 1  1  1 FALSE    1     1
#> 2  1  2 FALSE    3     3
#> 3  1  3 FALSE    6     6
#> 4  1  4 FALSE    0     0
#> 5  1  5  TRUE    1     1
#> 6  1  6 FALSE    7     7
#> 7  1  7  TRUE    0     0
#> 8  1  8  TRUE    1     1
#> 9  1  9 FALSE   10    10
#> 10 1 10  TRUE    0     0

^{Created on 2024-02-29 with reprex v2.0.2}