R 将日期时间间隔分配给日期时间事件

我有一个大于1M的个人日期时间检测间隔表（即在此期间连续检测到一个人）和一个表示闸门关闭时间的日期时间间隔表。单个检测分为“进入”或“退出”

使用%内的%我可以确定检测是否在任何间隔内

然而，我想做一些不同的事情。对于每个闸门关闭间隔，我想知道有多少人在外面被检测到，有多少人在里面被检测到。我相信最简单的方法是将每个检测事件分配给一个门周期，但是如果没有非常混乱的嵌套ifelse语句，我无法理解如何编写该函数

预期输出（非真实数据）：

闸门关闭间隔

new("Interval", .Data = c(-81060, -117060, -59400, -16200, -76680, 
-51000, -81120), start = structure(c(1412238660, 1412362800, 
1412434800, 1412454600, 1412542980, 1412602200, 1412690400), class = c("POSIXct", 
"POSIXt"), tzone = "UTC"), tzone = "UTC")

---

因此，您的示例数据存在一些问题，使得回答您的问题变得困难

您的示例检测数据包含许多不需要解决或理解您的问题的额外信息

您的示例间隔是向后的。（它们在开始之前就结束了）

您的示例数据没有以易于使用的格式设置（您不包括读取它所需的包）

也就是说，让我们简化并纠正这些问题：

library(tidyverse)  # needed because your data is in tibble format
library(lubridate)  # needed for time interval class used in your example

# I called your example data "Detections" and "Intervals", not importing them here

set.seed(914)
row.sample <- c(1,5,2, sample(1:100, 10)) # to include the samples not in the given intervals mixed with good values
use.cols <- c(1,4,6,7,9)  
Detections.use <- Detections[row.sample, use.cols]

Detections.use
# A tibble: 13 x 5
   Tag     site  species      StartDateTime_UTC   EndDateTime_UTC    
   <fct>   <fct> <chr>        <dttm>              <dttm>             
 1 5004.24 IC1   Striped Bass 2014-09-29 22:40:40 2014-09-29 22:46:35
 2 5004.24 RGD1  Striped Bass 2014-10-02 00:15:47 2014-10-02 00:30:08
 3 5004.24 IC1   Striped Bass 2014-09-29 22:49:15 2014-09-29 22:50:05
 4 5004.24 RGD1  Striped Bass 2014-10-04 18:38:40 2014-10-04 18:52:30
 5 5004.24 RGD1  Striped Bass 2014-10-05 07:15:43 2014-10-05 08:05:47
 6 5004.24 RGD1  Striped Bass 2014-10-03 02:21:36 2014-10-03 02:24:01
 7 5004.24 RGD1  Striped Bass 2014-10-03 12:13:46 2014-10-03 12:18:02
 8 5004.24 RGD1  Striped Bass 2014-10-03 09:30:15 2014-10-03 10:14:19
 9 5004.24 RGD1  Striped Bass 2014-10-04 14:09:39 2014-10-04 14:09:45
10 5004.24 RGD1  Striped Bass 2014-10-03 02:17:05 2014-10-03 02:18:05
11 5004.24 RGD1  Striped Bass 2014-10-05 03:08:26 2014-10-05 03:12:31
12 5004.24 RGD1  Striped Bass 2014-10-03 01:42:02 2014-10-03 01:58:18
13 5004.24 RGD1  Striped Bass 2014-10-02 00:33:12 2014-10-02 01:10:21


# Convert your intervals to a table, then correct them to make them positive time intevals
Interval.table <- tibble(Intervals, "ID" = seq(from = 1, to = length(Intervals))) %>%
      mutate(end_time = Intervals$start,
             start_time = end_time+Intervals$.Data,
             new_Interval = start_time %--% end_time) %>%
      select(ID, new_Interval)
Interval.table
# A tibble: 7 x 2
     ID new_Interval                                    
  <int> <Interval>                                      
1     1 2014-10-01 10:00:00 UTC--2014-10-02 08:31:00 UTC
2     2 2014-10-02 10:29:00 UTC--2014-10-03 19:00:00 UTC
3     3 2014-10-03 22:30:00 UTC--2014-10-04 15:00:00 UTC
4     4 2014-10-04 16:00:00 UTC--2014-10-04 20:30:00 UTC
5     5 2014-10-04 23:45:00 UTC--2014-10-05 21:03:00 UTC
6     6 2014-10-05 23:20:00 UTC--2014-10-06 13:30:00 UTC
7     7 2014-10-06 15:28:00 UTC--2014-10-07 14:00:00 UTC

编辑以添加

sapply

解决方案---

这也可以通过稍微平滑一点的

sapply

来完成，并且不会留下瑕疵：

which.join <- function(x, y) {
      z <- which(x %within% y)
      z <- ifelse(isTRUE(z > 0), z, NA)
      z
}

Detections.use$Interval <- sapply(Detections.use$StartDateTime_UTC, 
                             function(x) which.join(x,Interval.table$new_Interval)

---

which.join您能否提供一个示例，说明您正在寻找的输出内容（可以是虚拟数据），但格式可能会有所帮助。运行示例数据代码时，我还收到一个错误“getClass（Class，其中=topenv（parent.frame（）））：“Period”不是已定义的类”。也许可以将其剥离，使其更加精简—听起来您可以使用数据库连接方法将观测数据与间隔数据连接起来。这假设它们在两个单独的表中。@BrianFisher我已经更新了问题并修复了大门关闭期间的dput。我认为您应该能够使用类似
的东西（Detections$StartDateTime\u UTC%in%Interval）
来获取索引，然后您可以将其转换为ID。不过，我无法测试它，因为您提供的两组数据没有重叠（间隔时间都在2013年3月/4月，而检测时间都在2014年9月/10月）@BrianFisher这就是我尝试对数据进行子集的原因。不幸的是，它只返回整数0。我将尝试为该个体获得正确的间隔这与我得出的答案非常接近：
for（I in 1:length（dat_overlap$StartDateTime_UTC））{tryCatch（{dat_overlap$interval.id[I]
library(tidyverse)  # needed because your data is in tibble format
library(lubridate)  # needed for time interval class used in your example

# I called your example data "Detections" and "Intervals", not importing them here

set.seed(914)
row.sample <- c(1,5,2, sample(1:100, 10)) # to include the samples not in the given intervals mixed with good values
use.cols <- c(1,4,6,7,9)  
Detections.use <- Detections[row.sample, use.cols]

Detections.use
# A tibble: 13 x 5
   Tag     site  species      StartDateTime_UTC   EndDateTime_UTC    
   <fct>   <fct> <chr>        <dttm>              <dttm>             
 1 5004.24 IC1   Striped Bass 2014-09-29 22:40:40 2014-09-29 22:46:35
 2 5004.24 RGD1  Striped Bass 2014-10-02 00:15:47 2014-10-02 00:30:08
 3 5004.24 IC1   Striped Bass 2014-09-29 22:49:15 2014-09-29 22:50:05
 4 5004.24 RGD1  Striped Bass 2014-10-04 18:38:40 2014-10-04 18:52:30
 5 5004.24 RGD1  Striped Bass 2014-10-05 07:15:43 2014-10-05 08:05:47
 6 5004.24 RGD1  Striped Bass 2014-10-03 02:21:36 2014-10-03 02:24:01
 7 5004.24 RGD1  Striped Bass 2014-10-03 12:13:46 2014-10-03 12:18:02
 8 5004.24 RGD1  Striped Bass 2014-10-03 09:30:15 2014-10-03 10:14:19
 9 5004.24 RGD1  Striped Bass 2014-10-04 14:09:39 2014-10-04 14:09:45
10 5004.24 RGD1  Striped Bass 2014-10-03 02:17:05 2014-10-03 02:18:05
11 5004.24 RGD1  Striped Bass 2014-10-05 03:08:26 2014-10-05 03:12:31
12 5004.24 RGD1  Striped Bass 2014-10-03 01:42:02 2014-10-03 01:58:18
13 5004.24 RGD1  Striped Bass 2014-10-02 00:33:12 2014-10-02 01:10:21


# Convert your intervals to a table, then correct them to make them positive time intevals
Interval.table <- tibble(Intervals, "ID" = seq(from = 1, to = length(Intervals))) %>%
      mutate(end_time = Intervals$start,
             start_time = end_time+Intervals$.Data,
             new_Interval = start_time %--% end_time) %>%
      select(ID, new_Interval)
Interval.table
# A tibble: 7 x 2
     ID new_Interval                                    
  <int> <Interval>                                      
1     1 2014-10-01 10:00:00 UTC--2014-10-02 08:31:00 UTC
2     2 2014-10-02 10:29:00 UTC--2014-10-03 19:00:00 UTC
3     3 2014-10-03 22:30:00 UTC--2014-10-04 15:00:00 UTC
4     4 2014-10-04 16:00:00 UTC--2014-10-04 20:30:00 UTC
5     5 2014-10-04 23:45:00 UTC--2014-10-05 21:03:00 UTC
6     6 2014-10-05 23:20:00 UTC--2014-10-06 13:30:00 UTC
7     7 2014-10-06 15:28:00 UTC--2014-10-07 14:00:00 UTC

y <- c()
for (i in 1:length(Detections.use$StartDateTime_UTC)){
z<- which(Detections.use$StartDateTime_UTC[i] %within% Interval.table$new_Interval)
y[i] <- ifelse(isTRUE(z>0), z, NA)
}
Detections.use$Interval <- y
Detections.use
# A tibble: 13 x 6
   Tag     site  species      StartDateTime_UTC   EndDateTime_UTC     Interval
   <fct>   <fct> <chr>        <dttm>              <dttm>                 <int>
 1 5004.24 IC1   Striped Bass 2014-09-29 22:40:40 2014-09-29 22:46:35       NA
 2 5004.24 RGD1  Striped Bass 2014-10-02 00:15:47 2014-10-02 00:30:08        1
 3 5004.24 IC1   Striped Bass 2014-09-29 22:49:15 2014-09-29 22:50:05       NA
 4 5004.24 RGD1  Striped Bass 2014-10-04 18:38:40 2014-10-04 18:52:30        4
 5 5004.24 RGD1  Striped Bass 2014-10-05 07:15:43 2014-10-05 08:05:47        5
 6 5004.24 RGD1  Striped Bass 2014-10-03 02:21:36 2014-10-03 02:24:01        2
 7 5004.24 RGD1  Striped Bass 2014-10-03 12:13:46 2014-10-03 12:18:02        2
 8 5004.24 RGD1  Striped Bass 2014-10-03 09:30:15 2014-10-03 10:14:19        2
 9 5004.24 RGD1  Striped Bass 2014-10-04 14:09:39 2014-10-04 14:09:45        3
10 5004.24 RGD1  Striped Bass 2014-10-03 02:17:05 2014-10-03 02:18:05        2
11 5004.24 RGD1  Striped Bass 2014-10-05 03:08:26 2014-10-05 03:12:31        5
12 5004.24 RGD1  Striped Bass 2014-10-03 01:42:02 2014-10-03 01:58:18        2
13 5004.24 RGD1  Striped Bass 2014-10-02 00:33:12 2014-10-02 01:10:21        1

which.join <- function(x, y) {
      z <- which(x %within% y)
      z <- ifelse(isTRUE(z > 0), z, NA)
      z
}

Detections.use$Interval <- sapply(Detections.use$StartDateTime_UTC, 
                             function(x) which.join(x,Interval.table$new_Interval)