R 如何根据条件和上面的行操作行的值？

我有一个df，我想按

链接对它进行分组

，并按

时间对它进行排序

。然后，每次

type=='vehicle leave traffic'

列

count

中的单元格应加上前一行单元格值的+1。如果

type==“车辆进入交通”

则应从前一行中扣除1

为了澄清，不应更改前一行的值，而应基于前一行的值更改该行的值

这是我的方法，但我得到的是0、1和2。我期望一些链接的值更高

parking_min <- cars %>% 
  group_by(link)%>% 
  dplyr::mutate(count = if_else(type == 'vehicle leaves traffic', lag(count, n=1,order_by=time)+1,lag(count))) %>%
  dplyr::mutate(count = if_else(type == 'vehicle enters traffic',lag(count, n=1, order_by=time)-1,lag(count)))

可能的产出：

          time                   type vehicle_id link count
18798  23707.0 vehicle enters traffic    1267069   90     0 #start point
64777  31209.0 vehicle leaves traffic     810534   90     1 #+1
64783  31210.0 vehicle enters traffic     810534   90     0 #-1
90025  36230.0 vehicle leaves traffic      51825   90     1
90030  36231.0 vehicle enters traffic      51825   90     0
102868 38925.0 vehicle leaves traffic    1326473   90     1
105834 39583.0 vehicle leaves traffic    1199672   90     2 #here as well 1+1 =2
108690 40198.0 vehicle leaves traffic    1111105   90     3 #2+1 =3
111727 40818.0 vehicle enters traffic    1111105   90     2 #3-1 =2
118283 41974.0 vehicle leaves traffic     532654   90     3
124349 42895.0 vehicle enters traffic     532654   90     2
125700 43099.0 vehicle leaves traffic    1267069   90     3
129642 43683.0 vehicle enters traffic    1199672   90     2
135888 44645.0 vehicle leaves traffic    1398907   90     3
142577 45730.0 vehicle enters traffic    1398907   90     2
148772 46785.0 vehicle leaves traffic    1239391   90     3
161264 48846.0 vehicle enters traffic    1239391   90     2
161590 48905.0 vehicle enters traffic    1326473   90     1
182778 52790.0 vehicle leaves traffic      46491   90     2

---

最终我想找到每个链接的最大计数。但这可以在另一个步骤中完成，不需要成为解决方案的一部分，也许这有助于澄清问题。

我认为这正是您想要的：

df %>% 
  group_by(link) %>%
  arrange(time) %>%
  mutate(
    adder = case_when(
      type == "vehicle leaves traffic" ~ 1,
      type == "vehicle enters traffic" ~ -1,
      TRUE ~ 0),
    count = count + cumsum(adder)
    ) %>% 
  select(-adder)

df %>%
    arrange(link, time) %>%
    group_by(link) %>%
    mutate(vehicles_entered_traffic = cumsum(type == "vehicle enters traffic")
           , vehicles_left_traffic = cumsum(type == "vehicle leaves traffic")
           , count = count[1] + vehicles_left_traffic - vehicles_entered_traffic)

给

      time                   type vehicle_id link count
1  23707.0 vehicle enters traffic    1267069   90     0
2  31209.0 vehicle leaves traffic     810534   90     1
3  31210.0 vehicle enters traffic     810534   90     0
4  36230.0 vehicle leaves traffic      51825   90     1
5  36231.0 vehicle enters traffic      51825   90     0
6  38925.0 vehicle leaves traffic    1326473   90     1
7  39583.0 vehicle leaves traffic    1199672   90     2
8  40198.0 vehicle leaves traffic    1111105   90     3
9  40818.0 vehicle enters traffic    1111105   90     2
10 41974.0 vehicle leaves traffic     532654   90     3

---

我想这就是你想要的：

df %>% 
  group_by(link) %>%
  arrange(time) %>%
  mutate(
    adder = case_when(
      type == "vehicle leaves traffic" ~ 1,
      type == "vehicle enters traffic" ~ -1,
      TRUE ~ 0),
    count = count + cumsum(adder)
    ) %>% 
  select(-adder)

df %>%
    arrange(link, time) %>%
    group_by(link) %>%
    mutate(vehicles_entered_traffic = cumsum(type == "vehicle enters traffic")
           , vehicles_left_traffic = cumsum(type == "vehicle leaves traffic")
           , count = count[1] + vehicles_left_traffic - vehicles_entered_traffic)

---

如果你能提供一个最小的可复制的例子（可能还有预期的输出），那会很有帮助。@Georgery，我相信我提供的最小可复制的例子。我将尝试描述所需的输出，手工操作有点棘手。@Georgery添加了一个解决方案示例：）我的意思是，最好提供一个可能包含10行而不是300行的示例数据集@乔治亚哈，对不起，我总是担心给太少^ ^我会尝试更好地考虑最小的必要行数：）感谢提示和您的解决方案！您可能想按时间添加一些排序？