increment column if FALSE, stop incrementing if TRUE (tidyverse, dplyr, R)

Question

I have a table that looks like this:

data <- structure(list(group = c(0L, 0L, 1L, 2L), id = c("1", "2", "3", 
"4"), m = c("ac1", "ac1", "ac1", "me0"), together = c(FALSE, 
FALSE, TRUE, TRUE)), class = c("tbl_df", "tbl", "data.frame"), row.names = c(NA, 
-4L))

I would like to make a column group that increments when together == FALSE and stays at the current number when together == TRUE.

This is my desired output:

# A tibble: 4 x 4
  id    m     together group
  <chr> <chr> <lgl>    <dbl>
1 1     ac1   FALSE        0
2 2     ac1   FALSE        1
3 3     ac1   TRUE         2
4 4     me0   TRUE         2

I've already tried solutions like R increment by 1 for every change in value column and restart the counter and it's not giving me exactly what I want..

# A tibble: 4 x 4
# Groups:   group [3]
  id    m     together group
  <chr> <chr> <lgl>    <int>
1 1     ac1   FALSE        0
2 2     ac1   FALSE        0
3 3     ac1   TRUE         1
4 4     me0   TRUE         2

See I want group to read 0,1,2,2 in this case. Any ideas? Thank you so much.

@akrun there aren't any more rows - the pattern is to increment if FALSE and hold the current value if TRUE — Nick Camarda, Apr 24 '22 at 17:20
i see what your saying, maybe I'm not explaining it accurately enough; it should be `c(0,1,2,2,3,4)` — Nick Camarda, Apr 24 '22 at 17:23

score 2 · Accepted Answer · answered Apr 24 '22 at 17:24

2

Here's a tidy solution:

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
data <- structure(list(id = c("1", "2", "3", "4"), 
                       m = c("ac1", "ac1", "ac1", "me0"), 
                       together = c(FALSE, FALSE, TRUE, TRUE)), 
                  class = c("tbl_df", "tbl", "data.frame"), row.names = c(NA, -4L))
data <- data %>% mutate(
  group = cumsum(c(0, na.omit(lag(!data$together)))))
data
#> # A tibble: 4 × 4
#>   id    m     together group
#>   <chr> <chr> <lgl>    <dbl>
#> 1 1     ac1   FALSE        0
#> 2 2     ac1   FALSE        1
#> 3 3     ac1   TRUE         2
#> 4 4     me0   TRUE         2

^{Created on 2022-04-24 by the reprex package (v2.0.1)}

answered Apr 24 '22 at 17:24

DaveArmstrong

18,377
2
13
25

@onyambu what from the output doesn't look like the desired output from above? Note, the desired output is the first block of output, not the second. – DaveArmstrong Apr 24 '22 at 17:48
eg try `c(FALSE, FALSE, TRUE, TRUE, FALSE, TRUE)` which should give you `c(0,1,2,2,3,4)` – Onyambu Apr 24 '22 at 17:51
He got it - this matches the first block of output – Nick Camarda Apr 24 '22 at 17:51
@onyambu I'm not sure how you end up with a 4 if there are only three `FALSE` values, each is supposed to increment 1 starting from 0, shouldn't the largest number be 3? – DaveArmstrong Apr 24 '22 at 17:54
Each group of TRUES have their own ID, and each group of FALSE has id increament, starting from 0. That is what OP posted in the comments when the question was asked – Onyambu Apr 24 '22 at 17:55

score 0 · Answer 2 · answered Apr 24 '22 at 17:44

Another way to solve this is as given below

data %>%
       mutate(grp2 = c(0, cumsum(!na.omit(together * lag(together)))))
    # A tibble: 4 x 5
      group id    m     together  grp2
      <int> <chr> <chr> <lgl>    <dbl>
    1     0 1     ac1   FALSE        0
    2     0 2     ac1   FALSE        1
    3     1 3     ac1   TRUE         2
    4     2 4     me0   TRUE         2

Using the data given in the comments and running the same code:

data1 <- data.frame(together=c(FALSE, FALSE, TRUE, TRUE, FALSE, TRUE))
data1 %>%
   mutate(grp2 = c(0, cumsum(!na.omit(together * lag(together)))))
  together grp2
1    FALSE    0
2    FALSE    1
3     TRUE    2
4     TRUE    2
5    FALSE    3
6     TRUE    4

increment column if FALSE, stop incrementing if TRUE (tidyverse, dplyr, R)

2 Answers2