将连续日期折叠成一行

时间:2018-12-12 20:03:09

标签: r data-cleaning

我看到了很多有关如何将连续日期合并为一行的主题,并尝试了其中的少数几个主题(包括this并使用lead中的dplyr),但是到目前为止,还没有找不到专门回答我问题的线程。

这是我的数据:

df <- data.frame(
    id = c("A", "A", "A", "B", "B", "C", "C", "C"),
    start = as.Date(c("2013-05-21", "2014-03-17", "2014-12-12", "2009-03-08", 
                      "2011-07-30", "2008-10-07", "2009-11-21", "2010-12-01")),
    end = as.Date(c("2014-03-16", "2014-12-11", NA, "2011-07-14", 
                    NA, "2009-11-20", NA, NA)),
    status = c("expired", "expired", "active", "expired", 
               "active", "expired", "expired", "active")
    )

下面是我想要的输出:

id          start          end          status
A           2013-05-21     NA           active
B           2009-03-08     2011-07-14   expired
B           2011-07-30     NA           active
C           2008-10-07     NA           active 

所以我想做的事情有三点:

1)如果行是连续的,即结束日期+ 1是下一行的开始日期,我想将它们折叠为一行(如ID A中所示)

2)如果行不是连续的,即结束日期+ 1不是下一行的开始日期,我想将它们分开(如ID B)

3)如果“过期”行没有结束日期,我仍然希望将它们折叠为一行(如ID C)

任何帮助将不胜感激!

1 个答案:

答案 0 :(得分:1)

您可以选择类似的东西

library(tidyverse)

df %>%
  group_by(id) %>%
  mutate(
    end = if_else(is.na(end), lead(start), end),
    flag = if_else(start <= lag(end) + 1, 0, 1),
    flag = if_else(is.na(flag), 0, flag),
    group = cumsum(flag),
    flag = NULL
  ) %>%
  group_by(id, group) %>%
  mutate(
    start = first(start),
    end = last(end),
    status = last(status)
  ) %>% ungroup() %>% 
  distinct(id, start, end, status)

输出:

# A tibble: 4 x 4
  id    start      end        status 
  <fct> <date>     <date>     <fct>  
1 A     2013-05-21 NA         active 
2 B     2009-03-08 2011-07-14 expired
3 B     2011-07-30 NA         active 
4 C     2008-10-07 NA         active