按累计和分组

Question

我想按累计金额分组。在我的小例子中，列值中的前两个数字总和恰好为 100，属于 A 组。接下来的三个数字总和小于 100 并转到 B 组，而最后一个数字超过 100 并得到自己的组 C -等等。

input <- data.frame(id=c(1:6),
                   value =c(99, 1,  33, 33, 33, 150))
input

desired_output <- data.frame(id=c(1:6),
              value =c(99, 1,  33, 33, 33, 150),
              group= c("A", "A", "B", "B", "B", "C"))

desired_output

谢谢

Answer 1

desired_output <- data.frame(id=c(1:6),
                             value =c(99, 1,  33, 33, 33, 150),
                             group= c("A", "A", "B", "B", "B", "C"))

library(tidyverse)

desired_output %>% 
  group_by(grp_new = cumsum(value - 1) %/% 100) %>% 
  mutate(grp_new = LETTERS[cur_group_id()]) %>% 
  ungroup() 
#> # A tibble: 6 x 4
#>      id value group grp_new
#>   <int> <dbl> <chr> <chr>  
#> 1     1    99 A     A      
#> 2     2     1 A     A      
#> 3     3    33 B     B      
#> 4     4    33 B     B      
#> 5     5    33 B     B      
#> 6     6   150 C     C

^{由 reprex package (v2.0.1)}

于 2022-04-15 创建

Answer 2

两个可能的 one-liners，purrr::accumulate 和 MESS::cumsumbinning：

purrr::accumulate

library(tidyverse)
group_by(input, grp = LETTERS[cumsum(value == accumulate(value, ~ ifelse(.x + .y <= 100, .x + .y, .y)))])

MESS::cumsumbinning

library(dplyr)
group_by(input, grp = LETTERS[MESS::cumsumbinning(value, 100)])

输出

# A tibble: 6 x 3
# Groups:   grp [3]
     id value grp  
  <int> <dbl> <chr>
1     1    99 A    
2     2     1 A    
3     3    33 B    
4     4    33 B    
5     5    33 B    
6     6   150 C

按累计和分组

Group according to cumulative sums

grouping

r

cumsum

dplyr