Question

我的代码

df <- df %>% group_by(user_id, groupInsideUID = cumsum(time)) %>%
 mutate(Rank = ifelse(row_number() == 1, 'New','Repeat'))

我的数据框示例如下：

id  user_id   groupInsideUID      time            Rank
30    11       0                  NA              New
31    11       1              2/1/19 7:35 PM      New
54    5        1              3/1/18 2:35 PM      New
322   5        2              7/3/18 2:50 PM      New
21    5        2                   NA             Repeat
13    5        3              8/3/18 2:50 PM      New
2445  2        0                   NA             New
111   2        0                   NA             Repeat
287   2        1              5/3/18 2:50 PM      New
221   14       0                   NA             New
2345  7        0                   NA             New

我想根据最短时间在每个user_id组中创建一个新的列Rank（新建和重复）。在每个user_id组中，任何id在最短时间之前应有一个时间，该时间应为New（包括NA），在最短时间之后将被重复（包括NA）。

结果应如下所示：

id  user_id        time            Rank
30    11            NA             New
31    11       2/1/19 7:35 PM      New
54    5        3/1/18 2:35 PM      New
322   5        7/3/18 2:50 PM      Repeat
21    5             NA             Repeat
13    5        8/3/18 2:50 PM      Repeat
2445  2             NA             New
111   2             NA             New
287   2        5/3/18 2:50 PM      New
221   14            NA             New
2345  7             NA             New

感谢任何提示！

Answer 1

一种选择是将“时间”转换为按“ user_id”分组的DateTime类，根据最小“时间”的位置创建“等级”

library(lubridate)
library(dplyr)
df %>%
  mutate(time = dmy_hm(time)) %>%
  group_by(user_id) %>%
 mutate(Rank = if(all(is.na(time))) "New" 
    else case_when(row_number() <= which.min(time) ~ "New",
      TRUE ~ "Repeat"))
# A tibble: 11 x 5
# Groups:   user_id [5]
#      id user_id groupInsideUID time                Rank  
#   <int>   <int>          <int> <dttm>              <chr> 
# 1    30      11              0 NA                  New   
# 2    31      11              1 2019-01-02 19:35:00 New   
# 3    54       5              1 2018-01-03 14:35:00 New   
# 4   322       5              2 2018-03-07 14:50:00 Repeat
# 5    21       5              2 NA                  Repeat
# 6    13       5              3 2018-03-08 14:50:00 Repeat
# 7  2445       2              0 NA                  New   
# 8   111       2              0 NA                  New   
# 9   287       2              1 2018-03-05 14:50:00 New   
#10   221      14              0 NA                  New   
#11  2345       7              0 NA                  New

数据

df <- structure(list(id = c(30L, 31L, 54L, 322L, 21L, 13L, 2445L, 111L, 
287L, 221L, 2345L), user_id = c(11L, 11L, 5L, 5L, 5L, 5L, 2L, 
2L, 2L, 14L, 7L), groupInsideUID = c(0L, 1L, 1L, 2L, 2L, 3L, 
0L, 0L, 1L, 0L, 0L), time = c(NA, "2/1/19 7:35 PM", "3/1/18 2:35 PM", 
"7/3/18 2:50 PM", NA, "8/3/18 2:50 PM", NA, NA, "5/3/18 2:50 PM", 
NA, NA), Rank = c("New", "New", "New", "New", "Repeat", "New", 
"New", "Repeat", "New", "New", "New")), class = "data.frame",
row.names = c(NA, 
-11L))

如何根据R中的分组时间条件分配新列？

1 个答案:

数据