计算R中不规则时间间隔数据的总和

Question

我有两个数据框：数据框 1，它是具有分类值的不规则时间间隔数据，以及数据框 2，它是具有整数值的规则间隔数据。

数据框 1

Start Date	End Date	Category
1980-01-05	1983-02-17	A
1983-02-17	1987-01-02	B
1987-01-02	1989-11-10	C
1989-11-10	1992-03-20	D

数据框 2

Date	variable 1	variable 2	variable 3	...
1980-01-01	0	0	2	...
1980-02-01	0	0	0	...
1980-03-01	0	0	0	...
1980-04-01	0	1	2	...
1980-05-01	0	1	0	...
1980-06-01	-1	0	1	...
1980-07-01	-2	0	1	...
1980-08-01	-1	0	2	...
1980-09-01	0	2	1	...
1980-10-01	0	0	2	...
...	...	...	...	...

使用此数据框，我想将数据框 2 中的观察结果合并到数据框 1 中，将数据框 1 中开始日期到结束日期区间的值相加。

所以输出应该是这样的：

Start Date	End Date	Category	variable 1	variable 2	variable 3	...
1980-01-05	1983-02-17	A	Sum of variable 1 from the start date to end date	Sum of variable 2 from the start date to end date	Sum of variable 3 from the start date to end date	...
1983-02-17	1987-01-02	B	Sum of variable 1 from the start date to end date	Sum of variable 2 from the start date to end date	Sum of variable 3 from the start date to end date	...
1987-01-02	1989-11-10	C	Sum of variable 1 from the start date to end date	Sum of variable 2 from the start date to end date	Sum of variable 3 from the start date to end date	...
1989-11-10	1992-03-20	D	Sum of variable 1 from the start date to end date	Sum of variable 2 from the start date to end date	Sum of variable 3 from the start date to end date	...

Answer 1

这是使用我的 santoku 包的解决方案。你也可以使用cut.Date。

library(santoku)
library(dplyr)

# I assume that df1$Start_Date is a Date object
start_dates <- df1$Start_Date
df2$Start_Date <- santoku::chop(df2$Date, start_dates, labels = lbl_endpoint())
df2$Start_Date <- as.Date(as.character(df2$Start_Date))

df_summ <- df2 %>% 
             left_join(df1,  by = "Start_Date") %>%
             group_by(Start_Date) %>%
             summarize(
               End_Date   = End_Date[1],
               Category   = Category[1],
               Variable_1 = sum(Variable_1),
               Variable_2 = sum(Variable_2),
               Variable_3 = sum(Variable_3)
             )

您也可以将across() 用于变量 1 到 3，但上面的内容非常简单明了，但会重复一些。