我想创建一个新变量，该变量指示visit_date是否在ID列出的日期范围内
我已使用此代码进行逐行比较，但我想扩展此代码并将ID的所有行与该ID列出的所有间隔行进行比较
df <- df %>%
group_by(id) %>%
mutate(between_any = ifelse((visit_date >= start & visit_date <= end), 1))
我也尝试过创建一个间隔变量，并在变异前使用crossing（visit_date，interval），但是我无法对日期对象使用Crossing。
以下是一些示例数据：
df <- data.frame(id = c("a","a","a","a","a","b","b","b"),
visit_date = c("2001-08-22","2001-09-21","2001-10-30","2001-11-10","2001-12-20","2002-12-22", "2003-04-30","2003-05-10"),
start = c(NA,"2001-09-21",NA,"2001-11-10",NA,"2002-12-22", "2003-04-30",NA),
end = c(NA, "2001-11-01",NA,"2001-11-10",NA,"2002-12-22","2003-06-01",NA))
> df
id visit_date start end
a 2001-08-22 <NA> <NA>
a 2001-09-21 2001-09-21 2001-11-01
a 2001-10-30 <NA> <NA>
a 2001-11-10 2001-11-10 2001-11-10
a 2001-12-20 <NA> <NA>
b 2002-12-22 2002-12-22 2002-12-22
b 2003-04-30 2003-04-30 2003-06-01
b 2003-05-10 <NA> <NA>
我想要的输出如下：
id visit_date start end between_any
a 2001-08-22 <NA> <NA> 0
a 2001-09-21 2001-09-21 2001-11-01 1
a 2001-10-30 <NA> <NA> 1
a 2001-11-10 2001-11-10 2001-11-10 1
a 2001-12-20 <NA> <NA> 0
b 2002-12-22 2002-12-22 2002-12-22 1
b 2003-04-30 2003-04-30 2003-06-01 1
b 2003-05-10 <NA> <NA> 1
谢谢！
答案 0 :(得分：3)
in_range包中的 data.table函数可以做到这一点...
library(data.table)
df <- df %>%
group_by(id) %>%
mutate(between_any = as.numeric((inrange(visit_date, start, end))))
#> df
# id visit_date start end between_any
#1 a 2001-08-22 <NA> <NA> 0
#2 a 2001-09-21 2001-09-21 2001-11-01 1
#3 a 2001-10-30 <NA> <NA> 1
#4 a 2001-11-10 2001-11-10 2001-11-10 1
#5 a 2001-12-20 <NA> <NA> 0
#6 b 2002-12-22 2002-12-22 2002-12-22 1
#7 b 2003-04-30 2003-04-30 2003-06-01 1
#8 b 2003-05-10 <NA> <NA> 1
以data.table形式...
dt <- setDT(df)
dt[, between_any := inrange(visit_date, start, end),
by = id]
答案 1 :(得分：2)
我的回答不像我想要的那样“漂亮”，但是它可以带您找到想要去的地方。
我首先将您的日期转换为日期：
library(lubridate)
library(dplyr)
library(tibble)
library(tidyr)
library(purrr)
df <- data.frame(id = c("a","a","a","a","a","b","b","b"),
visit_date = c("2001-08-22","2001-09-21","2001-10-30","2001-11-10","2001-12-20","2002-12-22", "2003-04-30","2003-05-10"),
start = c(NA,"2001-09-21",NA,"2001-11-10",NA,"2002-12-22", "2003-04-30",NA),
end = c(NA, "2001-11-01",NA,"2001-11-10",NA,"2002-12-22","2003-06-01",NA)) %>%
mutate_at(-1,as.Date)
> df
id visit_date start end
1 a 2001-08-22 <NA> <NA>
2 a 2001-09-21 2001-09-21 2001-11-01
3 a 2001-10-30 <NA> <NA>
4 a 2001-11-10 2001-11-10 2001-11-10
5 a 2001-12-20 <NA> <NA>
6 b 2002-12-22 2002-12-22 2002-12-22
7 b 2003-04-30 2003-04-30 2003-06-01
8 b 2003-05-10 <NA> <NA>
接下来，我为每个组创建一个间隔列表：
df_intervals <- df %>%
mutate_at(-1,as.Date) %>%
filter(!is.na(start),
!is.na(end)) %>%
mutate(interval = start %--% end) %>%
select(id,interval) %>%
group_by(id)
> df_intervals
# A tibble: 4 x 2
# Groups: id [2]
id interval
<fct> <S4: Interval>
1 a 2001-09-21 UTC--2001-11-01 UTC
2 a 2001-11-10 UTC--2001-11-10 UTC
3 b 2002-12-22 UTC--2002-12-22 UTC
4 b 2003-04-30 UTC--2003-06-01 UTC
最后，我将间隔数据与基于id的原始数据结合起来，并在间隔内搜索visit_date。
df_output <- df %>% as.tbl() %>%
inner_join(df_intervals) %>%
mutate(between_any = map2_lgl(visit_date,interval,~ .x >= int_start(.y) & .x <= int_end(.y))) %>%
group_by(id,visit_date,start,end) %>%
summarise(between_any = as.numeric(any(between_any)))
> df_output
# A tibble: 8 x 5
# Groups: id, visit_date, start [8]
id visit_date start end between_any
<fct> <date> <date> <date> <dbl>
1 a 2001-08-22 NA NA 0
2 a 2001-09-21 2001-09-21 2001-11-01 1
3 a 2001-10-30 NA NA 1
4 a 2001-11-10 2001-11-10 2001-11-10 1
5 a 2001-12-20 NA NA 0
6 b 2002-12-22 2002-12-22 2002-12-22 1
7 b 2003-04-30 2003-04-30 2003-06-01 1
8 b 2003-05-10 NA NA 1
答案 2 :(得分：0)
另一种可能是：
df %>%
rowid_to_column() %>%
full_join(df %>%
filter(!is.na(start) & !is.na(end)) %>%
mutate(interval = interval(ymd(start), ymd(end))) %>%
select(id, interval), by = c("id" = "id")) %>%
group_by(rowid, id) %>%
summarise(between_any = max(ymd(visit_date) %within% interval * 1)) %>%
left_join(df %>%
rowid_to_column(), by = c("rowid" = "rowid",
"id" = "id")) %>%
ungroup() %>%
select(-rowid)
id between_any visit_date start end
<fct> <dbl> <fct> <fct> <fct>
1 a 0 2001-11-08 <NA> <NA>
2 a 1 2001-09-21 2001-09-21 2001-11-01
3 a 1 2001-10-30 <NA> <NA>
4 a 1 2001-11-10 2001-11-10 2001-11-10
5 a 0 2001-12-20 <NA> <NA>
6 b 1 2002-12-22 2002-12-22 2002-12-22
7 b 1 2003-04-30 2003-04-30 2003-06-01
8 b 1 2003-05-10 <NA> <NA>
在这里，首先创建interval变量，然后基于“ id”执行完全连接。其次，它检查“ visit_date”是否在每个“ id”和“ rowid”的间隔内。最后，它将结果与原始数据结合起来。