R检查并计算向量group_by中的字符串，考虑字符串的出现顺序

Question

数据采用以下格式，我必须使用日期对其进行分组。为方便起见，我将其显示为数字。

Msg <- c("Errors","Errors", "Start","Stop","Start","Stop","Errors","Errors","Start","Stop",
         "Stop" ,"Start","Errors","Start","Stop","Start" ,"Stop" ,
         "Errors", "Start","Errors","Stop", "Start", "LostControl","LostControl", "Errors",
         "Failed", "Stop", "Start","Failed","Stop","Stop","Start","Stop","Start","Error","Start",
         "Failed", "Stop")
Date <- c(11,11,11,11,11,11,11,12,12,12,12,12,12,14,14,14,14, 19,19,19,19,
        20,20,20,20,20,20,21,21,21,21,22,22,22,22,22,22,22)
data<- data.frame(Msg,Date)

我需要统计每个 START-STOP 周期中失败的次数，按日期汇总。
数据具有三种类型的消息。 Errors 和 Failed 是两种失败消息，而 LostControl 不是失败。 条件是 Failed msg 在该 START-STOP 循环中不应有 LostControl msg。如果它前面只有 Errors，它就是失败。 另外，如果只找到“Errors”消息，也不会被视为失败。

编辑：在 Msg 向量中，如果找到两个开始或停止，则 START_STOP 循环是从极端开始到极端停止。如果 START 后面没有 STOP，则将其忽略。

编辑一行添加为 - (Msg =Stop, Date=20)

Answer 1

我们可以修改我昨天在你的post 中写的那个函数。

between_valid_anchors <- function(x, bgn = "Start", end = "Stop") {
  are_anchors <- x %in% c(bgn, end)
  xid <- seq_along(x)
  id <- xid[are_anchors]
  x <- x[are_anchors]
  start_pos <- id[which(x == bgn & c("", head(x, -1L)) %in% c("", end))]
  stop_pos <- id[which(x == end & c(tail(x, -1L), "") %in% c("", bgn))]
  if (length(start_pos) < 1L || length(stop_pos) < 1L)
    return(logical(length(xid)))
  xid %in% unlist(mapply(`:`, start_pos, stop_pos))
}

那么就

library(dplyr)

data %>% 
  group_by(Date) %>% 
  filter(between_valid_anchors(Msg)) %>% 
  summarise(Msg = sum(Msg %in% c("Err", "Errors", "Failed")))

输出

`summarise()` ungrouping output (override with `.groups` argument)
# A tibble: 6 x 2
   Date   Msg
  <dbl> <int>
1    11     0
2    12     0
3    14     0
4    19     1
5    21     1
6    22     2

更新

您可以再添加一个过滤器以仅选择感兴趣的消息（即 Start、Stop、Failed、LostControl）。然后，只需对所有 Msg == "Failed" 求和，而不是 lag(Msg) == "LostControl"

library(dplyr)

data %>% 
  group_by(Date) %>% 
  filter(between_valid_anchors(Msg)) %>% 
  filter(Msg %in% c("Start", "Stop", "Failed", "LostControl")) %>% 
  summarise(Msg = sum(Msg == "Failed" & lag(Msg, default = "") != "LostControl"))

输出

`summarise()` ungrouping output (override with `.groups` argument)
# A tibble: 7 x 2
   Date   Msg
  <dbl> <int>
1    11     0
2    12     0
3    14     0
4    19     0
5    20     0
6    21     1
7    22     1