计算日期之间的实例数

Question

假设我有以下数据集：

library(data.table)
library(lubridate)

store_DT <- data.table(date = seq.Date(from = as.Date("2019-10-01"),
                                       to = as.Date("2019-10-05"),
                                       by = "day"),
                       store = c(rep("A",5),rep("B",5)))

    date         store
 1: 2019-10-01     A
 2: 2019-10-02     A
 3: 2019-10-03     A
 4: 2019-10-04     A
 5: 2019-10-05     A
 6: 2019-10-01     B
 7: 2019-10-02     B
 8: 2019-10-03     B
 9: 2019-10-04     B
10: 2019-10-05     B

这只是一个 data.table 商店 x 日期观察。

假设我还有另一个 data.table 的员工开始和结束时间（含）：

roster_DT <- data.table(
  store = c("A", "A", "A", "A", "B", "B","B", "B"),
  employee_ID = 1:8,
  start_date = c("2019-09-30", "2019-10-02", "2019-10-03", "2019-10-04",
                 "2019-09-30", "2019-10-02", "2019-10-03", "2019-10-04"),
  end_date = c("2019-10-04", "2019-10-04", "2019-10-05", "2019-10-06",
               "2019-10-04", "2019-10-04", "2019-10-05", "2019-10-06")
)

   store employee_ID start_date   end_date
1:     A           1 2019-09-30 2019-10-04
2:     A           2 2019-10-02 2019-10-04
3:     A           3 2019-10-03 2019-10-05
4:     A           4 2019-10-04 2019-10-06
5:     B           5 2019-09-30 2019-10-04
6:     B           6 2019-10-02 2019-10-04
7:     B           7 2019-10-03 2019-10-05
8:     B           8 2019-10-04 2019-10-06

我想做的只是计算每个商店在任何给定日期的员工人数，然后将其带回 store_DT。这里的复杂之处在于 roster_DT 指定了日期范围。现在，一种解决方案是使用建议 here 简单地扩展 roster_DT。但是实际的数据集比较大，扩容不是efficient/feasible。所以我想知道是否还有其他方法。

我要查找的最终数据集是：

    date         store   employees
 1: 2019-10-01     A      1
 2: 2019-10-02     A      2
 3: 2019-10-03     A      3
 4: 2019-10-04     A      4
 5: 2019-10-05     A      2
 6: 2019-10-01     B      1
 7: 2019-10-02     B      2
 8: 2019-10-03     B      3
 9: 2019-10-04     B      4
10: 2019-10-05     B      2

我的数据集中有很多商店和很多员工，所以我希望有一个 data.table 解决方案。

非常感谢！

Answer 1

请在下面找到使用 lubridate 库和 data.table 库的 foverlaps() 函数的解决方案（reprex）。

Reprex

代码

library(data.table)
library(lubridate)

# Convert 'start_date' and 'end_date' columns into class 'date'
sel_cols <- c("start_date", "end_date")
roster_DT[, (sel_cols) := lapply(.SD, ymd), .SDcols = sel_cols]

# Create a dummy variable in the data.table 'store_DT'
store_DT[, dummy := date]

# Set keys for the data.table 'roster_DT'
setkey(roster_DT, start_date, end_date)

# Merge the two data.tables with 'foverlaps()' and summarize the resulting data.table to get the requested data.table (i.e. 'Results')
Results <- foverlaps(store_DT,roster_DT, by.x=c("date", "dummy"), type = "within")[, dummy := NULL][,.(employees = .N/2), by = .(date, store)][]

# Reorder the data.table 'Results' by 'store', then 'date'         
setorder(Results, store, date)

-输出

Results
#>           date store employees
#>  1: 2019-10-01     A         1
#>  2: 2019-10-02     A         2
#>  3: 2019-10-03     A         3
#>  4: 2019-10-04     A         4
#>  5: 2019-10-05     A         2
#>  6: 2019-10-01     B         1
#>  7: 2019-10-02     B         2
#>  8: 2019-10-03     B         3
#>  9: 2019-10-04     B         4
#> 10: 2019-10-05     B         2

^{由 reprex package (v2.0.1)}

于 2021-11-17 创建

计算日期之间的实例数

Counting the number of instances between dates

r

data.table