Conditional replace in vector-CodePudding

I am trying to replace all values with NAs in a vector (by group) if there are at least two values that are greater than 4 where x is between 2 and 3. In this example, in group a, there are 2 values greater than 4 for 2 <= x <= 3.

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

tibble(
  grp = c("a", "a", "a", "b", "b", "b"),
  x = c(1, 2, 3, 1, 2, 3),
  val = c(4, 5, 6, 1, 2, 1)
) %>%
  group_by(grp) %>%
  mutate(val2 = ifelse(sum(val[between(x, 2, 3)] > 4) >= 2, NA, val))
#> # A tibble: 6 × 4
#> # Groups:   grp [2]
#>   grp       x   val  val2
#>   <chr> <dbl> <dbl> <dbl>
#> 1 a         1     4    NA
#> 2 a         2     5    NA
#> 3 a         3     6    NA
#> 4 b         1     1     1
#> 5 b         2     2     1
#> 6 b         3     1     1

Expected output

tibble(
  grp = c("a", "a", "a", "b", "b", "b"),
  x = c(1, 2, 3, 1, 2, 3),
  val = c(4, 5, 6, 1, 2, 1),
  val2 = c(NA, NA, NA, 1, 2, 1)
)
#> # A tibble: 6 × 4
#>   grp       x   val  val2
#>   <chr> <dbl> <dbl> <dbl>
#> 1 a         1     4    NA
#> 2 a         2     5    NA
#> 3 a         3     6    NA
#> 4 b         1     1     1
#> 5 b         2     2     2
#> 6 b         3     1     1

^{Created on 2021-10-25 by the reprex package (v2.0.1)}

CodePudding user response：

The problem is that ifelse return a vector with length equal to the first parameter. Since sum(val[between(x, 2, 3)] > 4) >= 2 returns a logical vector of length 1, only the first val is returned and then it's recycled to the full length. For example ifelse(TRUE, 1:3, 11:13) will only return 1. You could either use rep to repeat that value for the full length

mutate(val2 = ifelse(rep(sum(val[between(x, 2, 3)] > 4) >= 2, n()), NA, val))

or use a standard if/else statement

mutate(val2 = if(sum(val[between(x, 2, 3)] > 4) >= 2) NA else val)