【发布时间】：2016-11-18 08:47:30
【问题描述】：
我有这个样本
data <- structure(list(id = structure(c(1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L,
2L, 2L), .Label = c("a", "b"), class = "factor"), minTime = structure(c(1L,
2L, 3L, 4L, 5L, 7L, 6L, 8L, 9L, 10L), .Label = c("2014-06-06 07:39:50",
"2014-06-07 02:24:32", "2014-06-07 15:14:29", "2014-06-07 2:29",
"2014-06-08 5:40", "2014-06-18 17:54:42", "2014-06-18 2:45",
"2014-06-19 02:37:53", "2014-06-19 19:15", "2014-06-19 22:15"
), class = "factor"), maxTime = structure(c(1L, 3L, 4L, 2L, 5L,
6L, 7L, 8L, 9L, 10L), .Label = c("2014-06-07 01:41:31", "2014-06-07 10:01",
"2014-06-07 14:44:08", "2014-06-07 22:31:02", "2014-06-08 5:50",
"2014-06-18 2:50", "2014-06-19 01:49:05", "2014-06-19 18:51:36",
"2014-06-19 20:15", "2014-06-19 23:15"), class = "factor"), duration.minutes = c(NA,
740L, 437L, 452L, NA, NA, 474L, 974L, 4062L, 353L), event = structure(c(1L,
4L, 4L, 2L, 2L, 1L, 4L, 4L, 3L, 4L), .Label = c("enter", "exit",
"stop", "trip"), class = "factor")), .Names = c("id", "minTime",
"maxTime", "duration.minutes", "event"), class = "data.frame", row.names = c(NA,
-10L))
如果事件“旅行”对于每个 ID 都是连续的，我想合并它们。
id a 和 b 的连续行程已合并：
- maxTime 已编辑为包含行程的完整长度
- duration.minutes 也已编辑。
例如，这个输出：
dput(output.wanted)
structure(list(id = structure(c(1L, 1L, 1L, 2L, 2L, 1L, 2L), .Label = c("a",
"b"), class = "factor"), minTime = structure(c(5L, 6L, 7L, 2L,
1L, 3L, 4L), .Label = c("18.6.2014 17:54", "18.6.2014 2:45",
"19.6.2014 19:15", "19.6.2014 22:15", "6.6.2014 7:39", "7.6.2014 2:24",
"7.6.2014 2:29"), class = "factor"), maxTime = structure(c(5L,
7L, 6L, 1L, 2L, 3L, 4L), .Label = c("18.6.2014 2:50", "19.6.2014 18:51",
"19.6.2014 20:15", "19.6.2014 23:15", "7.6.2014 1:41", "7.6.2014 10:01",
"7.6.2014 22:31"), class = "factor"), duration.minutes = c(NA,
1177L, 452L, NA, 1448L, 4062L, 353L), event = structure(c(1L,
4L, 2L, 1L, 4L, 3L, 4L), .Label = c("enter", "exit", "stop",
"trip"), class = "factor")), .Names = c("id", "minTime", "maxTime",
"duration.minutes", "event"), class = "data.frame", row.names = c(NA,
-7L))
关键是我希望合并所有事件（行程或停止），如果它们对于相同的 ID 是连续的。但如果停靠点或行程是唯一的，则保持不变
我一直在尝试使用 group_by 和 mutate 来做到这一点，但我有点迷茫......
--
我找到了解决方案，但事件 stops
这是一个示例：
> dput(total)
structure(list(Ship = c(205482000, 205482000, 205482000, 205482000,
205482000, 205482000, 205482000, 205482000, 205482000, 205482000,
205482000, 205482000), minTime = structure(c(1401570241, 1401969219,
1401981860, 1402052108, 1402768362, 1402772602, 1402841443, 1402855773,
1403056361, 1403278916, 1403290856, 1403367735), class = c("POSIXct",
"POSIXt")), maxTime = structure(c(1401966639, 1401980399, 1402051849,
1402056430, 1402772313, 1402839873, 1402852433, 1403052550, 1403276355,
1403289496, 1403367596, 1403371285), class = c("POSIXct", "POSIXt"
)), duration.minutes = structure(c(6607, 186, 1166, NA, NA, 1121,
183, 3280, 3667, 176, 1279, NA), class = "difftime", units = "mins"),
event = c("stop", "trip", "trip", "exit", "enter", "trip",
"trip", "stop", "stop", "trip", "trip", "exit"), dist.sailed = c(NA,
50254.2034817555, 349194.108518887, NA, NA, 347816.081064252,
50035.8859874946, NA, NA, 49982.687612038, 351978.737678528,
NA)), .Names = c("Ship", "minTime", "maxTime", "duration.minutes",
"event", "dist.sailed"), class = "data.frame", row.names = c(4L,
5L, 6L, 2L, 1L, 7L, 8L, 9L, 10L, 11L, 12L, 3L))
以下代码不会产生停止的完整持续时间，而只会产生 NA：
total <- total %>%
group_by(Ship) %>%
mutate(new_id = data.table::rleid(event)) %>%
group_by(event, new_id, Ship) %>%
mutate(duration.minutes = ifelse(event == 'trip', sum(duration.minutes), duration.minutes), maxTime = tail(maxTime, 1))%>%
mutate(duration.minutes = ifelse(event == 'stop', sum(duration.minutes), duration.minutes), maxTime = tail(maxTime, 1))%>%
mutate(dist.sailed = ifelse(event == 'trip', sum(dist.sailed), dist.sailed), dist.sailed = tail(dist.sailed, 1)) %>%
filter(!duplicated(duration.minutes)) %>%
select(-new_id)
我在@Sotos 的代码中添加了mutate(duration.minutes = ifelse(event == 'stop', sum(duration.minutes), duration.minutes), maxTime = tail(maxTime, 1))%>%）
【问题讨论】：
-
请提供一个最小的例子还是我们需要了解所有的列/行？
-
所有 col/rows 更容易理解：有事件（行程、停止、进入或退出）。 minTime 是发泄的开始，maxTime 是结束。
-
这个问题不是和这个类似吗？ stackoverflow.com/questions/38054518/… 除非您需要按 id 和事件分组？
-
@theArun，是的，我尝试使用这个并在那里添加了评论。他们找到的解决方案不起作用...
-
@弗洛尼。确实如此，我写了解决方案。我刚刚测试过它仍然有效:)