将日期聚合为 R 中的日期间隔/周期

Question

我有以下示例数据：

require(tibble)
sample_data <- tibble(
                      emp_name = c("john", "john", "john", "john","john","john", "john"), 
                      task = c("carpenter", "carpenter","carpenter", "painter", "painter", "carpenter", "carpenter"),
                      date_stamp = c("2019-01-01","2019-01-02", "2019-01-03", "2019-01-07", "2019-01-08", "2019-01-30", "2019-02-02")
                      )

为此，我需要根据日期聚合成间隔。

规则是：如果为同一属性列出的下一个date_stamp之间没有日期，那么它应该被聚合。 否则， date_stamp_from和date_stamp_to应该等于date_stamp 。

desired_result <- tibble(
                  emp_name = c("john", "john","john", "john"),
                  task = c("carpenter","painter", "carpenter", "carpenter"),
                  date_stamp_from = c("2019-01-01","2019-01-07", "2019-01-30", "2019-02-02"),
                  date_stamp_to = c("2019-01-03","2019-01-08", "2019-01-30", "2019-02-02"),
                  count_dates = c(3,2,1,1)
)

解决这个问题的最有效方法是什么？ 原始数据集大约有 10000 条记录。

Answer 1

我们可以使用diff和cumsum创建组并计算每个组中的first 、 last date_stamp和行数。

library(dplyr)

sample_data %>%
     mutate(date_stamp = as.Date(date_stamp)) %>%
     group_by(gr = cumsum(c(TRUE, diff(date_stamp) > 1))) %>%
     mutate(date_stamp_from = first(date_stamp), 
            date_stamp_to = last(date_stamp), 
            count_dates = n()) %>%
     slice(1L) %>%
     ungroup() %>%
     select(-gr, -date_stamp)

# A tibble: 4 x 5
#  emp_name task      date_stamp_from date_stamp_to count_dates
#  <chr>    <chr>     <date>          <date>              <int>
#1 john     carpenter 2019-01-01      2019-01-03              3
#2 john     painter   2019-01-07      2019-01-08              2
#3 john     carpenter 2019-01-30      2019-01-30              1
#4 john     carpenter 2019-02-02      2019-02-02              1