如何在R中以字符形式构建条件滚动序列

Question

I have a data frame with 4 columns

g represents some group id in the data

x represents some character type id on which I have to perform computation

date is date of action, to be used for some other analysis. (however, I have retained the same so that the order of records may be preserved)

action is action to be performed. This column has two values 'add' and 'reduce'. These are equal in number for each group g . Means if there are 5 records having 'add' then there will be 5 'reduce' in the same group. recorded as an ordered factor in data frame.

Basically, I have to create a new column by the following rules-

• Since first record in each group will always be 'add' therefore exact value of x may be used here

• In second and onward record, in each group, the value of x may be pasted to previous record given that 'action' value is 'add'

• In second and onward record, in each group, the value of x may be removed from previous record given that 'action' value is 'reduce'

dput of sample data is

df <- structure(list(g = c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 
3L, 3L, 4L, 4L, 4L, 4L, 5L, 5L, 5L, 5L, 6L, 6L, 7L, 7L, 8L, 8L, 
8L, 8L), x = c("1_", "1_", "2_", "2_", "2_", "1_", "2_", "1_", 
"1_", "1_", "1_", "2_", "2_", "2_", "1_", "2_", "3_", "3_", "3_", 
"3_", "4_", "4_", "3_", "3_", "3_", "3_", "3_", "3_", "4_", "5_", 
"4_", "5_", "6_", "6_", "6_", "6_", "7_", "7_", "7_", "7_", "8_", 
"8_", "8_", "8_"), date = structure(c(18262, 18264, 18265, 18266, 
18271, 18275, 18275, 18276, 18277, 18279, 18280, 18283, 18286, 
18287, 18288, 18291, 18262, 18264, 18275, 18276, 18277, 18288, 
18275, 18283, 18291, 18297, 18301, 18309, 18366, 18374, 18375, 
18381, 18309, 18319, 18328, 18347, 18364, 18367, 18303, 18309, 
18328, 18341, 18341, 18344), class = "Date"), action = structure(c(1L, 
1L, 1L, 1L, 2L, 2L, 2L, 2L, 1L, 1L, 2L, 1L, 2L, 1L, 2L, 2L, 1L, 
1L, 2L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 1L, 2L, 2L, 1L, 
1L, 2L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L), .Label = c("add", 
"reduce"), class = c("ordered", "factor"))), row.names = c(NA, 
-44L), groups = structure(list(g = 1:8, .rows = structure(list(
    1:16, 17:22, 23:28, 29:32, 33:36, 37:38, 39:40, 41:44), ptype = integer(0), class = c("vctrs_list_of", 
"vctrs_vctr", "list"))), row.names = c(NA, -8L), class = c("tbl_df", 
"tbl", "data.frame"), .drop = TRUE), class = c("grouped_df", 
"tbl_df", "tbl", "data.frame"))

df
# A tibble: 44 x 4
# Groups:   g [8]
       g x     date       action
   <int> <chr> <date>     <ord> 
 1     1 1_    2020-01-01 add   
 2     1 1_    2020-01-03 add   
 3     1 2_    2020-01-04 add   
 4     1 2_    2020-01-05 add   
 5     1 2_    2020-01-10 reduce
 6     1 1_    2020-01-14 reduce
 7     1 2_    2020-01-14 reduce
 8     1 1_    2020-01-15 reduce
 9     1 1_    2020-01-16 add   
10     1 1_    2020-01-18 add   
# ... with 34 more rows

desired result may be like -

head(as_tibble(df3_r),12)
# A tibble: 12 x 6
       X     g x     date       action result  
   <int> <int> <chr> <chr>      <chr>  <chr>   
 1     1     1 1_    01-01-2020 add    1_      
 2     2     1 1_    03-01-2020 add    1_1_    
 3     3     1 2_    04-01-2020 add    1_1_2_  
 4     4     1 2_    05-01-2020 add    1_1_2_2_
 5     5     1 2_    10-01-2020 reduce 1_1_2_  
 6     6     1 1_    14-01-2020 reduce 1_2_    
 7     7     1 2_    14-01-2020 reduce 1_      
 8     8     1 1_    15-01-2020 reduce NA      
 9     9     1 1_    16-01-2020 add    1_      
10    10     1 1_    18-01-2020 add    1_1_    
11    11     1 1_    19-01-2020 reduce 1_      
12    12     1 2_    22-01-2020 add    1_2_ 

Thanks in advance.

Answer 1

You can use accumulate2 from the purrr package.

library(dplyr)
library(purrr)
df %>% 
  mutate(result = unlist(accumulate2(
    x, action[-1L], ~if (..3 == "add") paste0(., .y) else sub(.y, "", .)
  )))

Output

   g  x       date action   result
1  1 1_ 2020-01-01    add       1_
2  1 1_ 2020-01-03    add     1_1_
3  1 2_ 2020-01-04    add   1_1_2_
4  1 2_ 2020-01-05    add 1_1_2_2_
5  1 2_ 2020-01-10 reduce   1_1_2_
6  1 1_ 2020-01-14 reduce     1_2_
7  1 2_ 2020-01-14 reduce       1_
8  1 1_ 2020-01-15 reduce         
9  1 1_ 2020-01-16    add       1_
10 1 1_ 2020-01-18    add     1_1_
11 1 1_ 2020-01-19 reduce       1_
12 1 2_ 2020-01-22    add     1_2_
13 1 2_ 2020-01-25 reduce       1_
14 1 2_ 2020-01-26    add     1_2_
15 1 1_ 2020-01-27 reduce       2_
16 1 2_ 2020-01-30 reduce         
17 2 3_ 2020-01-01    add       3_
18 2 3_ 2020-01-03    add     3_3_
19 2 3_ 2020-01-14 reduce       3_
20 2 3_ 2020-01-15 reduce         
21 2 4_ 2020-01-16    add       4_
22 2 4_ 2020-01-27 reduce         
23 3 3_ 2020-01-14    add       3_
24 3 3_ 2020-01-22 reduce         
25 3 3_ 2020-01-30    add       3_
26 3 3_ 2020-02-05 reduce         
27 3 3_ 2020-02-09    add       3_
28 3 3_ 2020-02-17 reduce         
29 4 4_ 2020-04-14    add       4_
30 4 5_ 2020-04-22    add     4_5_
31 4 4_ 2020-04-23 reduce       5_
32 4 5_ 2020-04-29 reduce         
33 5 6_ 2020-02-17    add       6_
34 5 6_ 2020-02-27    add     6_6_
35 5 6_ 2020-03-07 reduce       6_
36 5 6_ 2020-03-26 reduce         
37 6 7_ 2020-04-12    add       7_
38 6 7_ 2020-04-15 reduce         
39 7 7_ 2020-02-11    add       7_
40 7 7_ 2020-02-17 reduce         
41 8 8_ 2020-03-07    add       8_
42 8 8_ 2020-03-20 reduce         
43 8 8_ 2020-03-20    add       8_
44 8 8_ 2020-03-23 reduce