(dplyr) 使用 mutate()、case_when() 和 which() 时出错

Question

在以下情况下需要有关错误/警告的帮助：

我有一个带有观察日期的数据框和一个此类日期的向量。我想在数据框中创建具有下一个和上一个观察日期的新列。日期向量创建为：

library(glue)

dates = c("201902",
          "201906",
          "201911",
          "202002")

dates = glue("{dates}01")

dates = dates%>%
        as.Date(format = "%Y%m%d")%>%
        sort()

然后我的数据框有这个由dates元素组成的列，称为Date。我希望它创建具有下一个和上一个日期的列，或者如果它是开始/结束则保持相同。我正在使用：

library(dplyr)

my_df = my_df%>%
mutate(First_date = (Date == dates[1]),
       Last_date = (Date == dates[length(dates)]),
       Prev_date = case_when(First_date ~ Date,
                                   TRUE ~ dates[which(dates == Date)-1]),
       Next_date = case_when(Last_date ~ Date,
                                  TRUE ~ dates[which(dates == Date)+1])

示例：如果我有一个包含以下列的数据框：

>my_df$Date
[1] "2019-02-01" "2019-06-01" "2019-11-01" "2020-02-01"

我希望它结束​​：

>my_df$First_date
[1] TRUE FALSE FALSE FALSE
>my_df$Last_date
[1] FALSE FALSE FALSE TRUE
>my_df$Prev_date
[1] "2019-02-01" "2019-02-01" "2019-06-01" "2019-11-01"
>my_df$Next_date
[1] "2019-06-01" "2019-11-01" "2020-02-01" "2020-02-01"

我使用的测试数据框有 6 行，它会抛出这个错误和警告：

Error: `TRUE ~ dates[which(dates == Date) + 1]` must be length 6 or one, not 2
Run `rlang::last_error()` to see where the error occurred.
In addition: Warning messages:
1: In `==.default`(dates, Date) :
  longer object length is not a multiple of shorter object length
2: In `==.default`(dates, Date) :
  longer object length is not a multiple of shorter object length

我认为这一定与在 case_when() 内调用 which() 函数有关，在 mutate() 内，但我还没有设法弄清楚它到底哪里出错了。

第一次来这里提问，如有错误请见谅！

Answer 1

好的，这应该可以。首先，你的df：

dates = c("201902",
          "201906",
          "201911",
          "202002")

dates = glue("{dates}01")

dates = dates%>%
        as.Date(format = "%Y%m%d")%>%
        sort()

my_df <- data.frame(Date = dates)

然后，使用shift 函数：

my_df <- my_df %>%
  mutate(First_date = ifelse(Date == dates[1], TRUE, FALSE),
         Last_date = ifelse(Date == dates[length(dates)], TRUE, FALSE),
         Prev_date = shift(dates, n = 1, fill = dates[1]),
         Next_date = shift(dates, n = -1, fill = dates[length(dates)]))

> my_df
        Date First_date Last_date  Prev_date  Next_date
1 2019-02-01       TRUE     FALSE 2019-02-01 2019-06-01
2 2019-06-01      FALSE     FALSE 2019-02-01 2019-11-01
3 2019-11-01      FALSE     FALSE 2019-06-01 2020-02-01
4 2020-02-01      FALSE      TRUE 2019-11-01 2020-02-01