计算员工工作时间的缺勤时间

Question

我在 R 中遇到了一个问题，试图计算员工的实际工作时间，我从工作时间中减去缺勤时间。

对于一个预定的一天，可以有多个缺勤时间。

一个预定日期的示例数据框：

row_num StartDate           EndDate             Absence_StartDate   Absence_EndDate
1       2019-11-13 14:30:00 2019-11-13 18:30:00 2019-11-13 15:45:00 2019-11-13 16:15:00
2       2019-11-13 14:30:00 2019-11-13 18:30:00 2019-11-13 08:15:00 2019-11-13 14:00:00
3       2019-11-13 14:30:00 2019-11-13 18:30:00 2019-11-13 15:30:00 2019-11-13 16:30:00
4       2019-11-13 14:30:00 2019-11-13 18:30:00 2019-11-13 08:00:00 2019-11-13 15:00:00

您可以从此行重现数据框：

df <- data.frame(StartDate = rep("2019-11-13 14:30:00", 4),
                 EndDate = rep("2019-11-13 18:30:00", 4),
                 Absence_StartDate = c("2019-11-13 15:45:00", "2019-11-13 08:15:00", "2019-11-13 15:30:00", "2019-11-13 08:00:00"),
                 Absence_EndDate = c("2019-11-13 16:15:00", "2019-11-13 14:00:00", "2019-11-13 16:30:00", "2019-11-13 15:00:00"))

我面临的主要问题是，其中一些缺席彼此重叠，例如 row_num 1 和 3。第一行对我没有用处，因为第三行与第一行重叠（距工作时间 -120 分钟）。第 2 行不影响工作时间，因为它超出了员工的计划，第 4 行与第 2 行重叠，它会影响工作时间（-15 分钟）。

本例中工作时间为240分钟，缺勤150分钟，所以实际工作时间为90分钟。

考虑到 StartDate 和 EndDate 可能不同（但同一天）并且可能有多次缺勤，如何编写代码来计算实际工作时间（有些可能根本不会影响工作时间）。

尝试使用 lubridate 库中的 intervals 并使用 for 循环，但无法获得结果。

谢谢！

====== 更新 =======

“@AnilGoyal”下面的代码在大多数情况下都可以正常工作。但是，有一些我试图解决但无法解决的问题。

这是一个不起作用的例子：

dput(df2)
structure(list(empid = c(1, 1, 1, 1, 1, 1, 1, 1), Date = structure(c(18213, 
18213, 18213, 18213, 18213, 18213, 18213, 18213), class = "Date"), 
    presence_start = structure(c(1573648200, 1573648200, 1573648200, 
    1573648200, 1573624800, 1573624800, 1573624800, 1573624800
    ), tzone = "", class = c("POSIXct", "POSIXt")), presence_end = structure(c(1573655400, 
    1573655400, 1573655400, 1573655400, 1573646400, 1573646400, 
    1573646400, 1573646400), tzone = "", class = c("POSIXct", 
    "POSIXt")), emprsn = c(1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L), absence_start = structure(list(
        sec = c(0, 0, 0, 0, 0, 0, 0, 0), min = c(15L, 15L, 30L, 
        0L, 15L, 15L, 30L, 0L), hour = c(15L, 8L, 14L, 8L, 15L, 
        8L, 14L, 8L), mday = c(13L, 13L, 13L, 13L, 13L, 13L, 
        13L, 13L), mon = c(10L, 10L, 10L, 10L, 10L, 10L, 10L, 
        10L), year = c(119L, 119L, 119L, 119L, 119L, 119L, 119L, 
        119L), wday = c(3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L), yday = c(316L, 
        316L, 316L, 316L, 316L, 316L, 316L, 316L), isdst = c(0L, 
        0L, 0L, 0L, 0L, 0L, 0L, 0L), zone = c("EET", "EET", "EET", 
        "EET", "EET", "EET", "EET", "EET"), gmtoff = c(7200L, 
        7200L, 7200L, 7200L, 7200L, 7200L, 7200L, 7200L)), tzone = c("", 
    "EET", "EEST"), class = c("POSIXlt", "POSIXt")), absence_end = structure(list(
        sec = c(0, 0, 0, 0, 0, 0, 0, 0), min = c(15L, 0L, 30L, 
        0L, 15L, 0L, 30L, 0L), hour = c(16L, 14L, 16L, 14L, 16L, 
        14L, 16L, 14L), mday = c(13L, 13L, 13L, 13L, 13L, 13L, 
        13L, 13L), mon = c(10L, 10L, 10L, 10L, 10L, 10L, 10L, 
        10L), year = c(119L, 119L, 119L, 119L, 119L, 119L, 119L, 
        119L), wday = c(3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L), yday = c(316L, 
        316L, 316L, 316L, 316L, 316L, 316L, 316L), isdst = c(0L, 
        0L, 0L, 0L, 0L, 0L, 0L, 0L), zone = c("EET", "EET", "EET", 
        "EET", "EET", "EET", "EET", "EET"), gmtoff = c(7200L, 
        7200L, 7200L, 7200L, 7200L, 7200L, 7200L, 7200L)), tzone = c("", 
    "EET", "EEST"), class = c("POSIXlt", "POSIXt"))), row.names = c(NA, 
-8L), class = "data.frame")

> df2
  empid       Date      presence_start        presence_end emprsn       absence_start         absence_end
1     1 2019-11-13 2019-11-13 18:00:00 2019-11-13 20:00:00      1 2019-11-13 15:15:00 2019-11-13 16:15:00
2     1 2019-11-13 2019-11-13 18:00:00 2019-11-13 20:00:00      2 2019-11-13 08:15:00 2019-11-13 14:00:00
3     1 2019-11-13 2019-11-13 18:00:00 2019-11-13 20:00:00      3 2019-11-13 14:30:00 2019-11-13 16:30:00
4     1 2019-11-13 2019-11-13 18:00:00 2019-11-13 20:00:00      4 2019-11-13 08:00:00 2019-11-13 14:00:00
5     1 2019-11-13 2019-11-13 11:30:00 2019-11-13 17:30:00      1 2019-11-13 15:15:00 2019-11-13 16:15:00
6     1 2019-11-13 2019-11-13 11:30:00 2019-11-13 17:30:00      2 2019-11-13 08:15:00 2019-11-13 14:00:00
7     1 2019-11-13 2019-11-13 11:30:00 2019-11-13 17:30:00      3 2019-11-13 14:30:00 2019-11-13 16:30:00
8     1 2019-11-13 2019-11-13 11:30:00 2019-11-13 17:30:00      4 2019-11-13 08:00:00 2019-11-13 14:00:00

下面的代码，经过一些修改以考虑几个员工的情况，后来注意到，一个员工一天的几个工作时间（例如：
从 2019 年 11 月 13 日 8:00 到 2019 年 11 月 13 日 14:00 和
2019-11-13 14:30 至 2019-11-13 16:30)

因此，每个工作时间都会重复缺勤，但据我所知，这应该不是问题。

在此示例中，我尝试将 presence_start 添加到 group_by，但 total_absence 两次都为 0，我注意到，两个工作时间只有一个d2=1 的实例，也许是问题所在？此示例的 total_absence 应为 6 小时和 2 小时。

Answer 1

由于场景已经修改，我提出备用tidyverse 策略。希望这行得通。

对示例数据进行了一些修改，因此之前的场景也包括在内。

> dput(df)
structure(list(empid = c("1", "1", "1", "1", "1", "1", "1", "1", 
"2", "2", "2", "2", "2"), presence_start = structure(list(sec = c(0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0), min = c(0L, 0L, 0L, 0L, 
30L, 30L, 30L, 30L, 30L, 30L, 30L, 30L, 30L), hour = c(18L, 18L, 
18L, 18L, 11L, 11L, 11L, 11L, 14L, 14L, 14L, 14L, 14L), mday = c(13L, 
13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L), 
    mon = c(10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 
    10L, 10L, 10L), year = c(119L, 119L, 119L, 119L, 119L, 119L, 
    119L, 119L, 119L, 119L, 119L, 119L, 119L), wday = c(3L, 3L, 
    3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L), yday = c(316L, 
    316L, 316L, 316L, 316L, 316L, 316L, 316L, 316L, 316L, 316L, 
    316L, 316L), isdst = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L), zone = c("IST", "IST", "IST", "IST", "IST", 
    "IST", "IST", "IST", "IST", "IST", "IST", "IST", "IST"), 
    gmtoff = c(NA_integer_, NA_integer_, NA_integer_, NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_, NA_integer_)), class = c("POSIXlt", 
"POSIXt")), presence_end = structure(list(sec = c(0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0), min = c(0L, 0L, 0L, 0L, 30L, 30L, 
30L, 30L, 30L, 30L, 30L, 30L, 30L), hour = c(20L, 20L, 20L, 20L, 
17L, 17L, 17L, 17L, 18L, 18L, 18L, 18L, 18L), mday = c(13L, 13L, 
13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L), mon = c(10L, 
10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L), 
    year = c(119L, 119L, 119L, 119L, 119L, 119L, 119L, 119L, 
    119L, 119L, 119L, 119L, 119L), wday = c(3L, 3L, 3L, 3L, 3L, 
    3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L), yday = c(316L, 316L, 316L, 
    316L, 316L, 316L, 316L, 316L, 316L, 316L, 316L, 316L, 316L
    ), isdst = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L), zone = c("IST", "IST", "IST", "IST", "IST", "IST", 
    "IST", "IST", "IST", "IST", "IST", "IST", "IST"), gmtoff = c(NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_, 
    NA_integer_, NA_integer_)), class = c("POSIXlt", "POSIXt"
)), absrsn = c("1", "2", "3", "4", "1", "2", "3", "4", "5", "6", 
"7", "8", "9"), absence_start = structure(list(sec = c(0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0), min = c(15L, 15L, 30L, 0L, 
15L, 15L, 30L, 0L, 45L, 15L, 30L, 0L, 15L), hour = c(15L, 8L, 
14L, 8L, 15L, 8L, 14L, 8L, 15L, 8L, 15L, 8L, 18L), mday = c(13L, 
13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L), 
    mon = c(10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 
    10L, 10L, 10L), year = c(119L, 119L, 119L, 119L, 119L, 119L, 
    119L, 119L, 119L, 119L, 119L, 119L, 119L), wday = c(3L, 3L, 
    3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L), yday = c(316L, 
    316L, 316L, 316L, 316L, 316L, 316L, 316L, 316L, 316L, 316L, 
    316L, 316L), isdst = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L), zone = c("IST", "IST", "IST", "IST", "IST", 
    "IST", "IST", "IST", "IST", "IST", "IST", "IST", "IST"), 
    gmtoff = c(NA_integer_, NA_integer_, NA_integer_, NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_, NA_integer_)), class = c("POSIXlt", 
"POSIXt")), absence_end = structure(list(sec = c(0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0), min = c(15L, 0L, 30L, 0L, 15L, 0L, 
30L, 0L, 15L, 0L, 30L, 0L, 0L), hour = c(16L, 14L, 16L, 14L, 
16L, 14L, 16L, 14L, 16L, 14L, 16L, 15L, 19L), mday = c(13L, 13L, 
13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L, 13L), mon = c(10L, 
10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L), 
    year = c(119L, 119L, 119L, 119L, 119L, 119L, 119L, 119L, 
    119L, 119L, 119L, 119L, 119L), wday = c(3L, 3L, 3L, 3L, 3L, 
    3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L), yday = c(316L, 316L, 316L, 
    316L, 316L, 316L, 316L, 316L, 316L, 316L, 316L, 316L, 316L
    ), isdst = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L), zone = c("IST", "IST", "IST", "IST", "IST", "IST", 
    "IST", "IST", "IST", "IST", "IST", "IST", "IST"), gmtoff = c(NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_, 
    NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_, 
    NA_integer_, NA_integer_)), class = c("POSIXlt", "POSIXt"
)), date = structure(c(18213, 18213, 18213, 18213, 18213, 18213, 
18213, 18213, 18213, 18213, 18213, 18213, 18213), class = "Date")), row.names = c(NA, 
-13L), spec = structure(list(cols = list(empid = structure(list(), class = c("collector_double", 
"collector")), presence_start = structure(list(), class = c("collector_character", 
"collector")), presence_end = structure(list(), class = c("collector_character", 
"collector")), absrsn = structure(list(), class = c("collector_double", 
"collector")), absence_start = structure(list(), class = c("collector_character", 
"collector")), absence_end = structure(list(), class = c("collector_character", 
"collector"))), default = structure(list(), class = c("collector_guess", 
"collector")), skip = 1), class = "col_spec"), class = c("spec_tbl_df", 
"tbl_df", "tbl", "data.frame"))

# A tibble: 13 x 7
   empid presence_start      presence_end        absrsn absence_start       absence_end         date      
   <chr> <dttm>              <dttm>              <chr>  <dttm>              <dttm>              <date>    
 1 1     2019-11-13 18:00:00 2019-11-13 20:00:00 1      2019-11-13 15:15:00 2019-11-13 16:15:00 2019-11-13
 2 1     2019-11-13 18:00:00 2019-11-13 20:00:00 2      2019-11-13 08:15:00 2019-11-13 14:00:00 2019-11-13
 3 1     2019-11-13 18:00:00 2019-11-13 20:00:00 3      2019-11-13 14:30:00 2019-11-13 16:30:00 2019-11-13
 4 1     2019-11-13 18:00:00 2019-11-13 20:00:00 4      2019-11-13 08:00:00 2019-11-13 14:00:00 2019-11-13
 5 1     2019-11-13 11:30:00 2019-11-13 17:30:00 1      2019-11-13 15:15:00 2019-11-13 16:15:00 2019-11-13
 6 1     2019-11-13 11:30:00 2019-11-13 17:30:00 2      2019-11-13 08:15:00 2019-11-13 14:00:00 2019-11-13
 7 1     2019-11-13 11:30:00 2019-11-13 17:30:00 3      2019-11-13 14:30:00 2019-11-13 16:30:00 2019-11-13
 8 1     2019-11-13 11:30:00 2019-11-13 17:30:00 4      2019-11-13 08:00:00 2019-11-13 14:00:00 2019-11-13
 9 2     2019-11-13 14:30:00 2019-11-13 18:30:00 5      2019-11-13 15:45:00 2019-11-13 16:15:00 2019-11-13
10 2     2019-11-13 14:30:00 2019-11-13 18:30:00 6      2019-11-13 08:15:00 2019-11-13 14:00:00 2019-11-13
11 2     2019-11-13 14:30:00 2019-11-13 18:30:00 7      2019-11-13 15:30:00 2019-11-13 16:30:00 2019-11-13
12 2     2019-11-13 14:30:00 2019-11-13 18:30:00 8      2019-11-13 08:00:00 2019-11-13 15:00:00 2019-11-13
13 2     2019-11-13 14:30:00 2019-11-13 18:30:00 9      2019-11-13 18:15:00 2019-11-13 19:00:00 2019-11-13

现在遵循这种方法（在加载所需的两个库之后

library(tidyverse)
library(lubridate)

#calculating both presence and absence times for each employee for each day

df %>% group_by(empid, date, presence_start) %>%
  mutate(absence_start = if_else(absence_start < presence_start, presence_start, absence_start),
         absence_end = if_else(absence_end > presence_end, presence_end, absence_end),
         absence_end = if_else(absence_end < absence_start, absence_start, absence_end)) %>%
  arrange(empid, date, presence_start, absence_start) %>%
  mutate(rowid = row_number()) %>%
  mutate(absence_start = if_else(rowid >1 & absence_start < lag(absence_end), lag(absence_end), absence_start),
         absence_end = if_else(absence_end < absence_start, absence_start, absence_end)) %>%
  mutate(presence_end = if_else(rowid >1 & presence_end == lag(presence_end), presence_start, presence_end)) %>%
  ungroup() %>%
  mutate(presence = difftime(presence_end, presence_start, units = "mins"),
         absence = difftime(absence_end, absence_start, units = "mins")) %>%
  group_by(empid, date) %>% summarise(sum(presence), sum(absence))

检查结果

# A tibble: 2 x 4
# Groups:   empid [2]
  empid date       `sum(presence)` `sum(absence)`
  <chr> <date>     <drtn>          <drtn>        
1 1     2019-11-13 480 mins        270 mins      
2 2     2019-11-13 240 mins        105 mins 

手动检查 - 对于员工 1 两次 出席 次（第 1-4 和 5-8 行）（120+360=480 分钟）对于员工 2（第 9 行-13)（240 分钟）

员工 1 的

缺勤 次（第 1-4 行 - 0 分钟，因为所有内容都是重复的；第 6 行和第 8 行重叠 - 净缺勤时间（11:30 到 14:00 - 150 分钟）；第 5 行和第 7 行再次重叠（净缺勤 1430 到 1630 - 120 分钟）- emp-1 的总时间为 270 分钟。对于 emp -2 - 第 8 行和第 10 行重叠（净缺勤 1430 到 1500 - 30 分钟）；第 9 行& 11 再次重叠（净 1530 到 1630 - 60 分钟），第 13 行超过存在（净 1815 到 1830 - 15 分钟）总计 -105 分钟。

全部符合代码结果

祝你好运

Answer 2

针对这个问题，我提出了一个计算总缺勤时间的解决方案，如下-

显然，数据框将由员工组成，因此我复制了给定的 df，用于新创建的 empno 字段上的两名员工。

使用的示例数据-（请注意，列名也进行了一些调整）

> dput(df)
structure(list(empno = c(1, 1, 1, 1, 2, 2, 2, 2), rownum = c(1L, 
2L, 3L, 4L, 1L, 2L, 3L, 4L), presence_start = c("2019-11-13 14:30:00", 
"2019-11-13 14:30:00", "2019-11-13 14:30:00", "2019-11-13 14:30:00", 
"2019-11-13 14:30:00", "2019-11-13 14:30:00", "2019-11-13 14:30:00", 
"2019-11-13 14:30:00"), presence_end = c("2019-11-13 18:30:00", 
"2019-11-13 18:30:00", "2019-11-13 18:30:00", "2019-11-13 18:30:00", 
"2019-11-13 18:30:00", "2019-11-13 18:30:00", "2019-11-13 18:30:00", 
"2019-11-13 18:30:00"), absence_start = c("2019-11-13 15:45:00", 
"2019-11-13 08:15:00", "2019-11-13 15:30:00", "2019-11-13 08:00:00", 
"2019-11-13 15:45:00", "2019-11-13 08:15:00", "2019-11-13 15:30:00", 
"2019-11-13 08:00:00"), absence_end = c("2019-11-13 16:15:00", 
"2019-11-13 14:00:00", "2019-11-13 16:30:00", "2019-11-13 15:00:00", 
"2019-11-13 16:15:00", "2019-11-13 14:00:00", "2019-11-13 16:30:00", 
"2019-11-13 15:00:00")), row.names = c(NA, -8L), class = "data.frame")


> df
  empno rownum      presence_start        presence_end       absence_start         absence_end
1     1      1 2019-11-13 14:30:00 2019-11-13 18:30:00 2019-11-13 15:45:00 2019-11-13 16:15:00
2     1      2 2019-11-13 14:30:00 2019-11-13 18:30:00 2019-11-13 08:15:00 2019-11-13 14:00:00
3     1      3 2019-11-13 14:30:00 2019-11-13 18:30:00 2019-11-13 15:30:00 2019-11-13 16:30:00
4     1      4 2019-11-13 14:30:00 2019-11-13 18:30:00 2019-11-13 08:00:00 2019-11-13 15:00:00
5     2      1 2019-11-13 14:30:00 2019-11-13 18:30:00 2019-11-13 15:45:00 2019-11-13 16:15:00
6     2      2 2019-11-13 14:30:00 2019-11-13 18:30:00 2019-11-13 08:15:00 2019-11-13 14:00:00
7     2      3 2019-11-13 14:30:00 2019-11-13 18:30:00 2019-11-13 15:30:00 2019-11-13 16:30:00
8     2      4 2019-11-13 14:30:00 2019-11-13 18:30:00 2019-11-13 08:00:00 2019-11-13 15:00:00

缺勤时间的计算需要创建一些虚拟字段，如下所述-

library(tidyverse)
library(lubridate)

df2 <- df %>% pivot_longer(cols = -c(rownum, empno), names_to = c(".value", "event"), names_sep = "_") %>% 
  mutate(presence = ymd_hms(presence),
         absence = ymd_hms(absence),
         event = factor(event, levels = c("start", "end"), ordered =T),
         absence_m = case_when(absence < presence & event == 'start' ~ presence,
                                absence > presence & event == 'end' ~ presence,
                                TRUE ~ absence),
         dummy = ifelse(event == 'start', 1, -1)) %>%
  group_by(empno, rownum) %>%
  mutate(absence_m = case_when(event == 'end' & lag(absence_m) > absence_m ~ lag(absence_m), 
                               TRUE ~ absence_m)) %>%
  group_by(empno) %>% arrange(empno, absence_m) %>%
  mutate(dummy = cumsum(dummy),
         d2 = case_when(event == 'end' & lag(absence_m) == absence_m ~ 0,
                        event == 'start' & lead(absence_m) == absence_m ~0,
                        event == "start" & dummy == 1 ~ 1,
                        event == "end" & dummy == 0 ~ 1,
                        TRUE ~ 0)) %>%
  filter(d2 == 1) %>%
  mutate(absence_time = case_when(event == 'end' ~ (absence_m - lag(absence_m)),
                                  TRUE ~ 0)) %>%
  mutate(absence_time = replace_na(absence_time, 0)) %>%
  summarise(total_absence = sum(absence_time))


> df2
# A tibble: 2 x 2
  empno total_absence
  <dbl> <drtn>       
1     1 90 mins      
2     2 90 mins

我将把总存在的计算留给你，这应该不难。不过，如果您有任何疑问，请随时提出。