R在tidyverse中创建多个日期范围[重复]

Question

我正在尝试找到一种方法来分离我的数据中的各种间隔，以便与 ID 关联的每一行都有它的最小和最大持续时间，按 NA 月份划分。

我的数据看起来像这样，但有 9 列和 275 行

df <- data.frame("ID" = c(1:5),
 "jan" = c("2020-01-01",NA, "2020-01-01", "2020-01-01", "2020-01-01"),
 "feb" = c("2020-02-01", "2020-02-01", NA, "2020-02-01", "2020-02-01"),
 "mar" = c("2020-03-01", "2020-03-01", NA, "2020-03-01", NA),
 "apr" = c(NA, "2020-04-01", NA, "2020-04-01", "2020-04-01"), 
 "may" = c("2020-05-01", "2020-05-01", NA ,NA, "2020-05-01"),
"jun" = c("2020-06-01", "2020-06-01", "2020-06-01", NA, NA)
)

理想情况下，列应该是这样的：

ID Start1        Stop1          Start2       Stop2
1  "2020-01-01" "2020-03-01"   "2020-05-01"  "2020-06-01"
....

编辑：我已经编辑了这个要求，因为 A）它被标记为重复，尽管重复问题只是切线相关，B）因为我真的在寻找一个 tidyverse 解决方案 - 这就是我得到的

p>

Answer 1

您可以获取长格式数据，在每个NA 值处创建一个新的开始和停止组。为每个组选择first 和last 日期值并获取宽格式数据。

library(dplyr)

df %>%
  tidyr::pivot_longer(cols = -ID) %>%
  group_by(ID, grp = cumsum(is.na(value))) %>%
  na.omit() %>%
  summarise(start = first(value), 
            stop = last(value)) %>%
  mutate(grp = row_number()) %>%
  pivot_wider(names_from = grp, values_from = c(start, stop)) %>%
  select(ID, order(readr::parse_number(names(.))))

#     ID start_1    stop_1     start_2    stop_2    
#  <int> <chr>      <chr>      <chr>      <chr>     
#1     1 2020-01-01 2020-03-01 2020-05-01 2020-06-01
#2     2 2020-02-01 2020-06-01 NA         NA        
#3     3 2020-01-01 2020-01-01 2020-06-01 2020-06-01
#4     4 2020-01-01 2020-04-01 NA         NA        
#5     5 2020-01-01 2020-02-01 2020-04-01 2020-05-01