根据日期

Question

我有一个带有StartDate和EndDate的输入数据框，格式为日期： input_df: C1 C2 StartDate EndDate A B 9/5/2019 12/14/2019 C D 4/12/2019 5/14/2019 E F 12/5/2019 12/15/2019 我正在尝试根据某些条件实现以下输出： -如果sys.date（）小于或等于EndDate，那么我想保留该行并用Year + 1添加另一行 -如果sys.Date（）大于EndDate，则从年份到2020年替换2019

所需的输出是： output_df: C1 C2 StartDate EndDate A B 9/5/2019 12/14/2019 A B 9/5/2020 12/14/2020 C D 4/12/2020 5/14/2020 E F 12/5/2019 12/15/2019 E F 12/5/2020 12/15/2020 我已经探究了split_rows和lubridate，但是不确定如何将if条件与那些函数结合在一起。数据框很大，我正在尝试避免for循环这样做？

Answer 1

一种选择是使用case_when在“ StartDate”，“ EndDate”列上增加一年，然后与原始数据集绑定

library(dplyr)
library(lubridate)
input_df %>%
    mutate_at(3:4, ~ mdy(.) %m+% years(1)) %>%
    bind_rows(input_df %>%
             mutate_at(3:4, mdy)) %>% 
    arrange_all() %>% 
    group_by(C1, C2) %>% 
    slice(if(first(EndDate) <= Sys.Date()) n() else row_number())
# A tibble: 5 x 4
# Groups:   C1, C2 [3]
#  C1    C2    StartDate  EndDate   
#  <chr> <chr> <date>     <date>    
#1 A     B     2019-09-05 2019-12-14
#2 A     B     2020-09-05 2020-12-14
#3 C     D     2020-04-12 2020-05-14
#4 E     F     2019-12-05 2019-12-15
#5 E     F     2020-12-05 2020-12-15

---

或者另一种选择是根据条件uncount扩展行，然后通过增加一年replace最后一行

library(tidyr)
input_df %>% 
   mutate_at(3:4, mdy) %>%
   mutate(n = 1 + (Sys.Date() <= EndDate)) %>% 
   uncount(n) %>% 
   group_by(C1, C2) %>% 
   mutate_at(vars(-group_cols()), ~ replace(., n(), .[n()] + years(1))) 
# A tibble: 5 x 4
# Groups:   C1, C2 [3]
#  C1    C2    StartDate  EndDate   
#  <chr> <chr> <date>     <date>    
#1 A     B     2019-09-05 2019-12-14
#2 A     B     2020-09-05 2020-12-14
#3 C     D     2020-04-12 2020-05-14
#4 E     F     2019-12-05 2019-12-15
#5 E     F     2020-12-05 2020-12-15

---

或使用base R

nm1 <- c('StartDate', 'EndDate')
input_df[nm1] <- lapply(input_df[nm1], as.Date, format = "%m/%d/%Y")
i1 <- Sys.Date() <= input_df$EndDate
lst1 <- lapply(input_df[i1, nm1], function(date) 
   do.call(c, lapply(date, seq, length.out = 2, by = '1 year')))
input_df2 <- input_df[rep(seq_len(nrow(input_df)), i1 + 1),]
input_df2[rep(i1, i1 +1), nm1] <- lst1

数据

input_df <- structure(list(C1 = c("A", "C", "E"), C2 = c("B", "D", "F"), 
    StartDate = c("9/5/2019", "4/12/2019", "12/5/2019"), EndDate = c("12/14/2019", 
    "5/14/2019", "12/15/2019")), class = "data.frame", row.names = c(NA, 
-3L))