尝试将不同年份的几天变成几周

Question

我有代表不同年份的 data.frames。每年都有一个日期列。我想在每个变量中创建一个变量，它将一年的前 7 天、后 7 天等分组。所以“2020-01-17”将在“01-15 到 01-21”中

#sample Data    
x1 <- data.frame(
    day=c("2020-02-21" ,"2020-01-19" ,"2020-01-30" ,"2020-01-17" ,"2020-02-18" ,"2020-02-31", "2020-02-21" ,"2020-01-02" ,"2020-01-28", "2020-02-27" ,"2020-02-29","2020-02-11" ,"2020-01-05", "2020-02-06", "2020-02-10", "2020-01-31" ,"2020-02-18"),
    one = 1 )
    
    x2 <- data.frame(
    day=c("2021-02-21" ,"2021-01-19" ,"2021-01-30" ,"2021-01-17" ,"2021-02-18" ,"2021-02-31", "2021-02-21" ,"2021-01-02" ,"2021-01-28", "2021-02-27" ,"2021-02-29","2021-02-11" ,"2021-01-05", "2021-02-06", "2021-02-10", "2021-01-31" ,"2021-02-18"),
    one = 1 )
 

我知道如何将天数转换为周数，但是如果我按照建议使用

format

删除年份（从 R 中的日期中删除年份）它会生成一个字符，然后我不能使用

cut

。

x2$day <- as.Date( x2$day , "%Y-%m-%d") 
x1$day <- as.Date( x1$day , "%Y-%m-%d") 
x1$day2 <- format( x1$day , "%m-%d")
class( x1$day2)

如果我不删除年份，那么同一日期会在不同的星期结束。

using the cut function "2020-02-21", and "2021-02-21" are in different weeks.  I want them to be in the same bin

cut(as.Date(x2$day), breaks="week")
cut(as.Date(x1$day), breaks="week")

Answer 1

如果这包括闰年和非闰年，那将不起作用，因为一年中的第 i 天可能属于不同的日期范围。我们能做的就是根据组号给组命名。让我们将一年的前 14 天编号为 0，接下来的 14 天编号为 1，依此类推。

transform(x1, g = as.POSIXlt(day, format = "%Y-%m-%d")$yday %/% 14)

给予：

          day one  g
1  2020-02-21   1  3
2  2020-01-19   1  1
3  2020-01-30   1  2
4  2020-01-17   1  1
5  2020-02-18   1  3
6  2020-02-31   1 NA
7  2020-02-21   1  3
8  2020-01-02   1  0
9  2020-01-28   1  1
10 2020-02-27   1  4
11 2020-02-29   1  4
12 2020-02-11   1  2
13 2020-01-05   1  0
14 2020-02-06   1  2
15 2020-02-10   1  2
16 2020-01-31   1  2
17 2020-02-18   1  3

Answer 2

你和

cut()

走在正确的道路上。 使用跨越全年的每周序列，可以使用

findInterval()

找到与每个日期匹配的序列索引（即周数）。

library(dplyr)
library(lubridate)
x1 <- data.frame(
  day=ymd(c("2020-02-21" ,"2020-01-19" ,"2020-01-30" ,"2020-01-17" ,"2020-02-18" ,"2020-02-31", "2020-02-21" ,"2020-01-02" ,"2020-01-28", "2020-02-27" ,"2020-02-29","2020-02-11" ,"2020-01-05", "2020-02-06", "2020-02-10", "2020-01-31" ,"2020-02-18")),
  one = 1 )
#> Warning: 1 failed to parse.

x2 <- data.frame(
  day=ymd(c("2021-02-21" ,"2021-01-19" ,"2021-01-30" ,"2021-01-17" ,"2021-02-18" ,"2021-02-31", "2021-02-21" ,"2021-01-02" ,"2021-01-28", "2021-02-27" ,"2021-02-29","2021-02-11" ,"2021-01-05", "2021-02-06", "2021-02-10", "2021-01-31" ,"2021-02-18")),
  one = 1 )
#> Warning: 2 failed to parse.

x1seq <- seq(dmy("01-01-2020"), dmy("31-12-2020"), by = "7 days")
x2seq <- seq(dmy("01-01-2021"), dmy("31-12-2021"), by = "7 days")
x1 %>%
  mutate(week_start = x1seq[findInterval(day, x1seq,
                                         rightmost.closed = FALSE, 
                                         left.open = FALSE, 
                                         all.inside = FALSE)])
#>           day one week_start
#> 1  2020-02-21   1 2020-02-19
#> 2  2020-01-19   1 2020-01-15
#> 3  2020-01-30   1 2020-01-29
#> 4  2020-01-17   1 2020-01-15
#> 5  2020-02-18   1 2020-02-12
#> 6        <NA>   1       <NA>
#> 7  2020-02-21   1 2020-02-19
#> 8  2020-01-02   1 2020-01-01
#> 9  2020-01-28   1 2020-01-22
#> 10 2020-02-27   1 2020-02-26
#> 11 2020-02-29   1 2020-02-26
#> 12 2020-02-11   1 2020-02-05
#> 13 2020-01-05   1 2020-01-01
#> 14 2020-02-06   1 2020-02-05
#> 15 2020-02-10   1 2020-02-05
#> 16 2020-01-31   1 2020-01-29
#> 17 2020-02-18   1 2020-02-12
x2 %>%
  mutate(week_start = x2seq[findInterval(day, x2seq,
                                         rightmost.closed = FALSE, 
                                         left.open = FALSE, 
                                         all.inside = FALSE)])
#>           day one week_start
#> 1  2021-02-21   1 2021-02-19
#> 2  2021-01-19   1 2021-01-15
#> 3  2021-01-30   1 2021-01-29
#> 4  2021-01-17   1 2021-01-15
#> 5  2021-02-18   1 2021-02-12
#> 6        <NA>   1       <NA>
#> 7  2021-02-21   1 2021-02-19
#> 8  2021-01-02   1 2021-01-01
#> 9  2021-01-28   1 2021-01-22
#> 10 2021-02-27   1 2021-02-26
#> 11       <NA>   1       <NA>
#> 12 2021-02-11   1 2021-02-05
#> 13 2021-01-05   1 2021-01-01
#> 14 2021-02-06   1 2021-02-05
#> 15 2021-02-10   1 2021-02-05
#> 16 2021-01-31   1 2021-01-29
#> 17 2021-02-18   1 2021-02-12

^{创建于 2023-05-17 与 reprex v2.0.2}