自满足日期条件以来计算行数

【问题标题】：Count rows since a date condition is met自满足日期条件以来计算行数
【发布时间】：2018-04-06 03:30:06
【问题描述】：

我有一个包含以下列的 R data.table 

Date    Time    X   Y   
5/9/2016    12:00:00 AM 30  125 
5/9/2016    12:30:00 AM 27  127 
5/9/2016    1:00:00 AM  30  133 
5/9/2016    1:30:00 AM  25  115 
5/9/2016    2:00:00 AM  26  116 
5/9/2016    3:00:00 AM  34  128 
5/9/2016    3:30:00 AM  34  120 
5/9/2016    4:00:00 AM  26  133 
5/9/2016    5:00:00 AM  35  119 
5/9/2016    5:30:00 AM  35  126 
5/9/2016    6:00:00 AM  35  118 
5/9/2016    6:30:00 AM  28  120 
5/9/2016    7:00:00 AM  29  123 
5/9/2016    7:30:00 AM  30  125 
5/9/2016    8:30:00 AM  35  116 
5/9/2016    9:00:00 AM  32  123 
5/9/2016    9:30:00 AM  33  135 
5/9/2016    10:00:00 AM 32  127 
5/9/2016    10:30:00 AM 26  122 
5/9/2016    11:00:00 AM 28  122 
5/9/2016    11:30:00 AM 31  129 
5/9/2016    12:00:00 PM 29  127 
5/9/2016    12:30:00 PM 26  120 
5/9/2016    1:00:00 PM  32  126 
5/9/2016    1:30:00 PM  26  117 
5/9/2016    2:30:00 PM  32  133 
5/9/2016    3:00:00 PM  31  120 
5/9/2016    3:30:00 PM  35  129 
5/9/2016    4:00:00 PM  30  121 
5/9/2016    4:30:00 PM  35  132 
5/9/2016    5:00:00 PM  27  129 
5/9/2016    5:30:00 PM  27  122 
5/9/2016    6:00:00 PM  28  121 
5/9/2016    6:30:00 PM  35  127 
5/9/2016    7:00:00 PM  34  132 
5/9/2016    7:30:00 PM  25  133 
5/9/2016    8:00:00 PM  34  119 
5/9/2016    8:30:00 PM  33  125 
5/9/2016    9:00:00 PM  29  120 
5/9/2016    9:30:00 PM  34  125 
5/9/2016    10:00:00 PM 31  118 
5/9/2016    10:30:00 PM 33  118 
5/9/2016    11:00:00 PM 28  115 
5/9/2016    11:30:00 PM 34  126 
5/10/2016   12:00:00 AM 34  131 
5/10/2016   12:30:00 AM 25  130 
5/10/2016   1:00:00 AM  28  129 
5/10/2016   1:30:00 AM  26  132

我想要如下输出 -

Date    Time    X   Y   Rowcount@6hours
    5/9/2016    12:00:00 AM 30  125 0
    5/9/2016    12:30:00 AM 27  127 1
    5/9/2016    1:00:00 AM  30  133 2
    5/9/2016    1:30:00 AM  25  115 3
    5/9/2016    2:00:00 AM  26  116 4
    5/9/2016    3:00:00 AM  34  128 5
    5/9/2016    3:30:00 AM  34  120 6
    5/9/2016    4:00:00 AM  26  133 7
    5/9/2016    5:00:00 AM  35  119 8
    5/9/2016    5:30:00 AM  35  126 9
    5/9/2016    6:00:00 AM  35  118 10
    5/9/2016    6:30:00 AM  28  120 10
    5/9/2016    7:00:00 AM  29  123 10
    5/9/2016    7:30:00 AM  30  125 10
    5/9/2016    8:30:00 AM  35  116 9
    5/9/2016    9:00:00 AM  32  123 10
    5/9/2016    9:30:00 AM  33  135 10
    5/9/2016    10:00:00 AM 32  127 10
    5/9/2016    10:30:00 AM 26  122 10
    5/9/2016    11:00:00 AM 28  122 11
    5/9/2016    11:30:00 AM 31  129 11
    5/9/2016    12:00:00 PM 29  127 11
    5/9/2016    12:30:00 PM 26  120 11
    5/9/2016    1:00:00 PM  32  126 11
    5/9/2016    1:30:00 PM  26  117 11
    5/9/2016    2:30:00 PM  32  133 11
    5/9/2016    3:00:00 PM  31  120 11
    5/9/2016    3:30:00 PM  35  129 11
    5/9/2016    4:00:00 PM  30  121 11
    5/9/2016    4:30:00 PM  35  132 11
    5/9/2016    5:00:00 PM  27  129 11
    5/9/2016    5:30:00 PM  27  122 11
    5/9/2016    6:00:00 PM  28  121 11
    5/9/2016    6:30:00 PM  35  127 11
    5/9/2016    7:00:00 PM  34  132 11
    5/9/2016    7:30:00 PM  25  133 11
    5/9/2016    8:00:00 PM  34  119 11
    5/9/2016    8:30:00 PM  33  125 11
    5/9/2016    9:00:00 PM  29  120 12
    5/9/2016    9:30:00 PM  34  125 12
    5/9/2016    10:00:00 PM 31  118 12
    5/9/2016    10:30:00 PM 33  118 12
    5/9/2016    11:00:00 PM 28  115 12
    5/9/2016    11:30:00 PM 34  126 12
    5/10/2016   12:00:00 AM 34  131 12
    5/10/2016   12:30:00 AM 25  130 12
    5/10/2016   1:00:00 AM  28  129 12
    5/10/2016   1:30:00 AM  26  132 12

所以我的目标是计算我在过去 n 小时内获得的行数（例如 n=6）并将其附加到每一行（我想使用该值来确定 @987654327 中的 width 参数@ 职能）。这里的数据粒度为 30 分钟，但它可以更改为秒级别（并且也可能是不均匀的！）并且可能会丢失行

我尝试使用window.zoo 和xts，但无法找到解决方案。我在另一个链接中发布了同样的问题，那里的回复在一天内有效，但不是跨天Count number of rows between two timestamps in R dataframe

【问题讨论】：

  • 您知道Date 中有2015 年的日期吗？ （第 3 行和第 10 行）
  • 请edit 您的问题并发布dput(<your_data.table>) 的输出。这将有助于节省那些愿意帮助您的人的时间。请另见minimal reproducible example。谢谢。
  • 我在 Excel 中编辑了输入和输出，以展示我的输入和输出应该是什么样子。很抱歉造成混淆。

标签： r data.table time-series


【解决方案1】：

使用：

# convert to a 'data.table' (if necessary)
# add a datetime & rowid variable
DT <- as.data.table(df)[, `:=` (DateTime = as.POSIXct(paste(Date, Time), format = '%m/%d/%Y %I:%M:%S%p'),
                                rid = .I)][]

# calculate the count with a rolling join
DT[, count6hr := DT[.(DateTime = DateTime - 6*60*60, rid = rid)
                    , on = 'DateTime'
                    , roll = -Inf
                    , mult = 'last'
                    , i.rid - rid]][]

给出：

         Date       Time  X   Y            DateTime rid count6hr
 1:  5/9/2016 12:00:00AM 30 125 2016-05-09 00:00:00   1        0
 2:  5/9/2016 12:30:00AM 27 127 2016-05-09 00:30:00   2        1
 3:  5/9/2016  1:00:00AM 30 133 2016-05-09 01:00:00   3        2
 4:  5/9/2016  1:30:00AM 25 115 2016-05-09 01:30:00   4        3
 5:  5/9/2016  2:00:00AM 26 116 2016-05-09 02:00:00   5        4
 6:  5/9/2016  3:00:00AM 34 128 2016-05-09 03:00:00   6        5
 7:  5/9/2016  3:30:00AM 34 120 2016-05-09 03:30:00   7        6
 8:  5/9/2016  4:00:00AM 26 133 2016-05-09 04:00:00   8        7
 9:  5/9/2016  5:00:00AM 35 119 2016-05-09 05:00:00   9        8
10:  5/9/2016  5:30:00AM 35 126 2016-05-09 05:30:00  10        9
11:  5/9/2016  6:00:00AM 35 118 2016-05-09 06:00:00  11       10
12:  5/9/2016  6:30:00AM 28 120 2016-05-09 06:30:00  12       10
13:  5/9/2016  7:00:00AM 29 123 2016-05-09 07:00:00  13       10
14:  5/9/2016  7:30:00AM 30 125 2016-05-09 07:30:00  14       10
15:  5/9/2016  8:30:00AM 35 116 2016-05-09 08:30:00  15        9
16:  5/9/2016  9:00:00AM 32 123 2016-05-09 09:00:00  16       10
17:  5/9/2016  9:30:00AM 33 135 2016-05-09 09:30:00  17       10
18:  5/9/2016 10:00:00AM 32 127 2016-05-09 10:00:00  18       10
19:  5/9/2016 10:30:00AM 26 122 2016-05-09 10:30:00  19       10
20:  5/9/2016 11:00:00AM 28 122 2016-05-09 11:00:00  20       11
21:  5/9/2016 11:30:00AM 31 129 2016-05-09 11:30:00  21       11
22:  5/9/2016 12:00:00PM 29 127 2016-05-09 12:00:00  22       11
23:  5/9/2016 12:30:00PM 26 120 2016-05-09 12:30:00  23       11
24:  5/9/2016  1:00:00PM 32 126 2016-05-09 13:00:00  24       11
25:  5/9/2016  1:30:00PM 26 117 2016-05-09 13:30:00  25       11
26:  5/9/2016  2:30:00PM 32 133 2016-05-09 14:30:00  26       11
27:  5/9/2016  3:00:00PM 31 120 2016-05-09 15:00:00  27       11
28:  5/9/2016  3:30:00PM 35 129 2016-05-09 15:30:00  28       11
29:  5/9/2016  4:00:00PM 30 121 2016-05-09 16:00:00  29       11
30:  5/9/2016  4:30:00PM 35 132 2016-05-09 16:30:00  30       11
31:  5/9/2016  5:00:00PM 27 129 2016-05-09 17:00:00  31       11
32:  5/9/2016  5:30:00PM 27 122 2016-05-09 17:30:00  32       11
33:  5/9/2016  6:00:00PM 28 121 2016-05-09 18:00:00  33       11
34:  5/9/2016  6:30:00PM 35 127 2016-05-09 18:30:00  34       11
35:  5/9/2016  7:00:00PM 34 132 2016-05-09 19:00:00  35       11
36:  5/9/2016  7:30:00PM 25 133 2016-05-09 19:30:00  36       11
37:  5/9/2016  8:00:00PM 34 119 2016-05-09 20:00:00  37       11
38:  5/9/2016  8:30:00PM 33 125 2016-05-09 20:30:00  38       12
39:  5/9/2016  9:00:00PM 29 120 2016-05-09 21:00:00  39       12
40:  5/9/2016  9:30:00PM 34 125 2016-05-09 21:30:00  40       12
41:  5/9/2016 10:00:00PM 31 118 2016-05-09 22:00:00  41       12
42:  5/9/2016 10:30:00PM 33 118 2016-05-09 22:30:00  42       12
43:  5/9/2016 11:00:00PM 28 115 2016-05-09 23:00:00  43       12
44:  5/9/2016 11:30:00PM 34 126 2016-05-09 23:30:00  44       12
45: 5/10/2016 12:00:00AM 34 131 2016-05-10 00:00:00  45       12
46: 5/10/2016 12:30:00AM 25 130 2016-05-10 00:30:00  46       12
47: 5/10/2016  1:00:00AM 28 129 2016-05-10 01:00:00  47       12
48: 5/10/2016  1:30:00AM 26 132 2016-05-10 01:30:00  48       12

【讨论】：

    【解决方案2】：

    这也可以使用非等值连接来解决：

    library(data.table)
as.data.table(df)[
  , ts :=  lubridate::mdy_hms(paste(Date, Time))][
    , cnt := .SD[.(t6 = ts - 6*60*60, t0 = ts), 
                 on = .(ts >= t6, ts < t0), .N, by = .EACHI]$N][]

             Date     Time ampm  X   Y                  ts cnt
 1:  5/9/2016 12:00:00   AM 30 125 2016-05-09 00:00:00   0
 2:  5/9/2016 12:30:00   AM 27 127 2016-05-09 00:30:00   1
 3:  5/9/2016  1:00:00   AM 30 133 2016-05-09 01:00:00   2
 4:  5/9/2016  1:30:00   AM 25 115 2016-05-09 01:30:00   3
 5:  5/9/2016  2:00:00   AM 26 116 2016-05-09 02:00:00   4
 6:  5/9/2016  3:00:00   AM 34 128 2016-05-09 03:00:00   5
 7:  5/9/2016  3:30:00   AM 34 120 2016-05-09 03:30:00   6
 8:  5/9/2016  4:00:00   AM 26 133 2016-05-09 04:00:00   7
 9:  5/9/2016  5:00:00   AM 35 119 2016-05-09 05:00:00   8
10:  5/9/2016  5:30:00   AM 35 126 2016-05-09 05:30:00   9
11:  5/9/2016  6:00:00   AM 35 118 2016-05-09 06:00:00  10
12:  5/9/2016  6:30:00   AM 28 120 2016-05-09 06:30:00  10
13:  5/9/2016  7:00:00   AM 29 123 2016-05-09 07:00:00  10
14:  5/9/2016  7:30:00   AM 30 125 2016-05-09 07:30:00  10
15:  5/9/2016  8:30:00   AM 35 116 2016-05-09 08:30:00   9
16:  5/9/2016  9:00:00   AM 32 123 2016-05-09 09:00:00  10
17:  5/9/2016  9:30:00   AM 33 135 2016-05-09 09:30:00  10
18:  5/9/2016 10:00:00   AM 32 127 2016-05-09 10:00:00  10
19:  5/9/2016 10:30:00   AM 26 122 2016-05-09 10:30:00  10
20:  5/9/2016 11:00:00   AM 28 122 2016-05-09 11:00:00  11
21:  5/9/2016 11:30:00   AM 31 129 2016-05-09 11:30:00  11
22:  5/9/2016 12:00:00   PM 29 127 2016-05-09 12:00:00  11
23:  5/9/2016 12:30:00   PM 26 120 2016-05-09 12:30:00  11
24:  5/9/2016  1:00:00   PM 32 126 2016-05-09 13:00:00  11
25:  5/9/2016  1:30:00   PM 26 117 2016-05-09 13:30:00  11
26:  5/9/2016  2:30:00   PM 32 133 2016-05-09 14:30:00  11
27:  5/9/2016  3:00:00   PM 31 120 2016-05-09 15:00:00  11
28:  5/9/2016  3:30:00   PM 35 129 2016-05-09 15:30:00  11
29:  5/9/2016  4:00:00   PM 30 121 2016-05-09 16:00:00  11
30:  5/9/2016  4:30:00   PM 35 132 2016-05-09 16:30:00  11
31:  5/9/2016  5:00:00   PM 27 129 2016-05-09 17:00:00  11
32:  5/9/2016  5:30:00   PM 27 122 2016-05-09 17:30:00  11
33:  5/9/2016  6:00:00   PM 28 121 2016-05-09 18:00:00  11
34:  5/9/2016  6:30:00   PM 35 127 2016-05-09 18:30:00  11
35:  5/9/2016  7:00:00   PM 34 132 2016-05-09 19:00:00  11
36:  5/9/2016  7:30:00   PM 25 133 2016-05-09 19:30:00  11
37:  5/9/2016  8:00:00   PM 34 119 2016-05-09 20:00:00  11
38:  5/9/2016  8:30:00   PM 33 125 2016-05-09 20:30:00  12
39:  5/9/2016  9:00:00   PM 29 120 2016-05-09 21:00:00  12
40:  5/9/2016  9:30:00   PM 34 125 2016-05-09 21:30:00  12
41:  5/9/2016 10:00:00   PM 31 118 2016-05-09 22:00:00  12
42:  5/9/2016 10:30:00   PM 33 118 2016-05-09 22:30:00  12
43:  5/9/2016 11:00:00   PM 28 115 2016-05-09 23:00:00  12
44:  5/9/2016 11:30:00   PM 34 126 2016-05-09 23:30:00  12
45: 5/10/2016 12:00:00   AM 34 131 2016-05-10 00:00:00  12
46: 5/10/2016 12:30:00   AM 25 130 2016-05-10 00:30:00  12
47: 5/10/2016  1:00:00   AM 28 129 2016-05-10 01:00:00  12
48: 5/10/2016  1:30:00   AM 26 132 2016-05-10 01:30:00  12
         Date     Time ampm  X   Y                  ts cnt

    【讨论】：

    • 也感谢您的帮助！
    猜你喜欢
    • 2019-04-30
    • 1970-01-01
    • 2013-10-14
    • 2015-03-21
    • 2021-12-24
    • 2020-08-29
    • 2021-02-05
    • 1970-01-01
    • 1970-01-01
    相关资源
    最近更新 更多
    热门标签
    Java Python linux javascript Mysql C# Docker 算法 前端 SpringBoot Redis Vue spring 设计模式 .net core .net kubernetes c++ 数据库 数据结构 大数据 js 机器学习 微服务 Android Go 程序员 面试 JVM ASP.net core 云原生 人工智能 后端 PHP git CSS golang k8s Nginx Django mybatis 深度学习 多线程 React 架构 devops 爬虫 云计算 Spring Boot LeetCode