R 使用条件将数据帧一分为二

我有一个数据框

mydata = data.table(MyTimes = as.POSIXct(c("2015-01-01 00:00:03","2015-01-01 00:00:04","2015-01-01 00:00:18","2015-01-01 00:00:48","2015-01-01 00:00:48","2015-01-01 00:00:54","2015-01-01 00:01:12","2015-01-01 00:01:45"),tz = "GMT"),othercol= c(1,2,3,4,5,6,7))


 mydata
               MyTimes othercol
1: 2015-01-01 00:00:03        1
2: 2015-01-01 00:00:04        2
3: 2015-01-01 00:00:18        3
4: 2015-01-01 00:00:48        4
5: 2015-01-01 00:00:48        5
6: 2015-01-01 00:00:54        6
7: 2015-01-01 00:01:12        7
8: 2015-01-01 00:01:45        1

数据按时间排序，我想将此数据帧分为两个数据帧，有两个条件：

如果可能的话，断裂应该发生在中间，

但是，在相同的秒数下接近中断的时间必须在相同的数据帧中

在这个例子中有8行，我想在中间分解它。每行4行，但注意00:00:48将在两个数据帧中，根据上面的第2点，这是不可能的。也就是说，当你突破时，你不能在同一秒内突破

所以这里的数据帧可能是

data frame 1:
                   MyTimes othercol
     2015-01-01 00:00:03        1
     2015-01-01 00:00:04        2
     2015-01-01 00:00:18        3
     2015-01-01 00:00:48        4
     2015-01-01 00:00:48        5

data frame 2:
     2015-01-01 00:00:54        6
     2015-01-01 00:01:12        7
     2015-01-01 00:01:45        1

也可以是这样的：

data frame1:
   2015-01-01 00:00:03        1
   2015-01-01 00:00:04        2
   2015-01-01 00:00:18        3

data frame2:
    2015-01-01 00:00:48        4
    2015-01-01 00:00:48        5
    2015-01-01 00:00:54        6
    2015-01-01 00:01:12        7
    2015-01-01 00:01:45        1

无论哪种方式，00:00:48都在同一数据帧中，这是怎么回事

mydata = data.table(MyTimes = as.POSIXct(c("2015-01-01 00:00:03","2015-01-01 00:00:04","2015-01-01 00:00:18","2015-01-01 00:00:48","2015-01-01 00:00:48","2015-01-01 00:00:54","2015-01-01 00:01:12","2015-01-01 00:01:45"),tz = "GMT"),othercol= c(1,2,3,4,5,6,7))


 mydata
               MyTimes othercol
1: 2015-01-01 00:00:03        1
2: 2015-01-01 00:00:04        2
3: 2015-01-01 00:00:18        3
4: 2015-01-01 00:00:48        4
5: 2015-01-01 00:00:48        5
6: 2015-01-01 00:00:54        6
7: 2015-01-01 00:01:12        7
8: 2015-01-01 00:01:45        1

split(mydata, as.numeric(mydata$MyTimes) < median(as.numeric(mydata$MyTimes)))
$`FALSE`
               MyTimes secondcol
1: 2015-01-01 00:00:48         4
2: 2015-01-01 00:00:48         5
3: 2015-01-01 00:00:54         6
4: 2015-01-01 00:01:12         7
5: 2015-01-01 00:01:45         8

$`TRUE`
               MyTimes secondcol
1: 2015-01-01 00:00:03         1
2: 2015-01-01 00:00:04         2
3: 2015-01-01 00:00:18         3

split（mydata，as.numeric（mydata$MyTimes）

不像@DatamineR的解决方案那样优雅，但使用游程编码的替代方案是

library(data.table)

mydata[, grp := rleid(MyTimes)]  ## put times into groups
split(mydata, mydata$grp >= ceiling(max(mydata$grp)/2))

$`FALSE`
               MyTimes othercol grp
1: 2015-01-01 00:00:03        1   1
2: 2015-01-01 00:00:04        2   2
3: 2015-01-01 00:00:18        3   3

$`TRUE`
               MyTimes othercol grp
1: 2015-01-01 00:00:48        4   4
2: 2015-01-01 00:00:48        5   4
3: 2015-01-01 00:00:54        6   5
4: 2015-01-01 00:01:12        7   6
5: 2015-01-01 00:01:45        8   7

---

您的示例数据生成警告。我正在研究一个

data.table

解决方案（

library（data.table）setDT（mydata）[，grp:=rleid（MyTimes）]拆分（mydata，mydata$grp==天花板（max（mydata$grp）/2））

），它在逻辑上与您的类似（但您的更简单、更优雅：）@DatamineR我在.Call（“Crbindlist”，l，use.names，fill）：“Crbindlist”未从当前名称空间（data.table）解析@user3022875我明天会看一看