r - 如何正确标记组中第一个序列的结尾？

Question

这是我正在使用的数据帧类型和所需的列输出的示例。

reprEx <- data.frame(id = c(1,1,1,1,1,1,1,2,2,2,2,3,3,3),
             stage1 = c("open","open","open","approved","approved","open","declined","open","open","open","declined","open","approved","declined"))

Desireddf <- data.frame(id = c(1,1,1,1,1,1,1,2,2,2,2,3,3,3),
             stage1 = c("open","open","open","approved","approved","open","declined","open","open","open","declined","open","approved","declined"),
             desiredResult = c(0,0,1,1,0,0,1,0,0,1,1,1,1,1))

我正在尝试使用 dplyr 来正确标记分组 id 内的阶段更改。批准和拒绝的标记很容易，因为我只需标记第一个出现的拒绝或批准的情况:

    reprExWrong <- reprEx %>% group_by(id,stage1) %>%
  mutate(desiredResult = ifelse(stage1 == last(stage1) & stage1 == "open",1,
                                ifelse(stage1 == first(stage1) & stage1 %in% c("approved","declined"),1,0)
                                )
  )

问题出在开放阶段。我只想在第一个打开序列在一组 id 内结束时应用一个标志。使用我现在拥有的代码，它选择组内的最后一个打开，即使它不是第一个打开序列的一部分。例如:

reprExWrong <- data.frame(id = c(1,1,1,1,1,1,1,2,2,2,2,3,3,3),
             stage1 = c("open","open","open","approved","approved","open","declined","open","open","open","declined","open","approved","declined"),
             notdesiredResult = c(0,0,0,1,0,1,1,0,0,1,1,1,1,1))

在这种情况下，在 id1 内，我需要标志来显示序列在批准出现之前到达最后一个打开位置，而不是在批准出现之后打开。如果连续的打开是 id 中的第一个连续，我只需要最后一次出现打开的行中的标志。抱歉造成任何困惑，我很乐意进一步详细说明。这只是为了正确识别用于录制目的的阶段转换

Answer 1

当然有一些不那么复杂的东西，但这可能会起作用:

library(dplyr)

reprEx %>%
  group_by(id) %>%
  mutate(idx = cumsum(stage1 != lag(stage1, default = first(stage1))),
         desiredResult = case_when(
           idx != lag(idx, default = first(idx)) & stage1 != 'open' ~ 1,
           lead(idx, default = last(idx)) != idx & stage1 == 'open' & idx == min(idx[stage1 == 'open']) ~ 1,
           TRUE ~ 0
         ),
         idx = NULL
  )

输出:

   id   stage1 desiredResult
1   1     open             0
2   1     open             0
3   1     open             1
4   1 approved             1
5   1 approved             0
6   1     open             0
7   1 declined             1
8   2     open             0
9   2     open             0
10  2     open             1
11  2 declined             1
12  3     open             1
13  3 approved             1
14  3 declined             1