组合指示标记开始和结束的 POSIXct 列

Question

我有我的主要数据 df，还有一个单独的数据框 df_marker，表示 start 和 end 的标记。我的 df 数据框是一秒一秒的，并且是 POSIXct 格式。我想将 df_marker 加入 df 并能够在 start 和 end 内的每一秒显示标记.

代表：

library(dplyr)

df <- tibble(
  date = as.POSIXct(c("2020-11-17 12:00:00", "2020-11-17 12:00:01", "2020-11-17 12:00:02", "2020-11-17 12:00:03", "2020-11-17 12:00:04", "2020-11-17 12:00:05"))
)

df
#> # A tibble: 6 x 1
#>   date               
#>   <dttm>             
#> 1 2020-11-17 12:00:00
#> 2 2020-11-17 12:00:01
#> 3 2020-11-17 12:00:02
#> 4 2020-11-17 12:00:03
#> 5 2020-11-17 12:00:04
#> 6 2020-11-17 12:00:05

df_marker <- tibble(
  start = as.POSIXct("2020-11-17 12:00:02"),
  end = as.POSIXct("2020-11-17 12:00:05"),
  marker = "marker_1"
)

df_marker
#> # A tibble: 1 x 3
#>   start               end                 marker  
#>   <dttm>              <dttm>              <chr>   
#> 1 2020-11-17 12:00:02 2020-11-17 12:00:05 marker_1

这是我的预期输出：

#> # A tibble: 6 x 2
#>   date                marker  
#>   <dttm>              <chr>   
#> 1 2020-11-17 12:00:00 <NA>    
#> 2 2020-11-17 12:00:01 <NA>    
#> 3 2020-11-17 12:00:02 marker_1
#> 4 2020-11-17 12:00:03 marker_1
#> 5 2020-11-17 12:00:04 marker_1
#> 6 2020-11-17 12:00:05 marker_1

有什么想法吗？

Answer 1

这个有用吗：

library(dplyr)
library(tidyr)
library(purrr)
df_marker %>% mutate(date = map2(start, end, `:`)) %>% unnest(date) %>% 
mutate(date = as.POSIXct(date, origin = '1970-01-01')) %>% select(3,4) %>% 
right_join(df) %>% arrange(date) %>% select(2,1)
Joining, by = "date"
# A tibble: 6 x 2
  date                marker  
  <dttm>              <chr>   
1 2020-11-17 12:00:00 NA      
2 2020-11-17 12:00:01 NA      
3 2020-11-17 12:00:02 marker_1
4 2020-11-17 12:00:03 marker_1
5 2020-11-17 12:00:04 marker_1
6 2020-11-17 12:00:05 marker_1

一分钟一分钟：

library(lubridate)
df
# A tibble: 6 x 1
  date               
  <dttm>             
1 2020-11-17 12:00:00
2 2020-11-17 12:01:01
3 2020-11-17 12:02:02
4 2020-11-17 12:03:03
5 2020-11-17 12:04:04
6 2020-11-17 12:05:05
df_marker
# A tibble: 1 x 3
  start               end                 marker  
  <dttm>              <dttm>              <chr>   
1 2020-11-17 12:01:02 2020-11-17 12:03:05 marker_1
df_marker %>% mutate(across(start:end, ~ round_date(., unit = '1 minutes'))) %>% 
mutate(date = map2(start, end, `:`)) %>% unnest(date) %>% 
mutate(date = as.POSIXct(date, origin = '1970-01-01')) %>% select(3,4) %>% 
right_join(df %>% mutate(date = round_date(date, unit = '1 minutes'))) %>% 
arrange(date) %>% select(2,1)
Joining, by = "date"
# A tibble: 6 x 2
  date                marker  
  <dttm>              <chr>   
1 2020-11-17 12:00:00 NA      
2 2020-11-17 12:01:00 marker_1
3 2020-11-17 12:02:00 marker_1
4 2020-11-17 12:03:00 marker_1
5 2020-11-17 12:04:00 NA      
6 2020-11-17 12:05:00 NA

Answer 2

作为 Karthik S 的出色答案的替代方案，您还可以尝试 interval() for df_marker 和稍后的 %within% 来检查 df 中的任何日期在此时间间隔内。

这两个函数来自lubridate包。

library(lubridate)

df_marker <- df_marker %>%
  mutate(interval = interval(start,end))

  start               end                 marker    interval
  <dttm>              <dttm>              <chr>     <interval>
1 2020-11-17 12:01:02 2020-11-17 12:03:05 marker_1  2020-11-17 12:01:00 CET--2020-11-17 12:03:00 CET

df <- df %>%
  mutate(marker = ifelse(date %within% df_marker$interval,
                         df_marker$marker, NA))

> df
# A tibble: 7 x 2
  date                marker  
  <dttm>              <chr>   
1 2020-11-17 12:00:00 NA      
2 2020-11-17 12:01:01 NA      
3 2020-11-17 12:02:02 marker_1
4 2020-11-17 12:03:03 marker_1
5 2020-11-17 12:04:04 NA      
6 2020-11-17 12:05:05 NA      
7 2020-11-17 12:06:06 NA

为了一分钟一分钟，从两个表中抓取 round_date() 间隔和日期列。此功能也来自 lubridate 包。

df_marker <- df_marker %>%
  mutate(interval = interval(round_date(start,"minute"), round_date(end,"minute")))

df <- df %>%
  mutate(marker = ifelse(round_date(date, "minute") %within% df_marker$interval,
                         df_marker$marker, NA))
> df
# A tibble: 7 x 2
  date                marker  
  <dttm>              <chr>   
1 2020-11-17 12:00:00 NA      
2 2020-11-17 12:01:01 marker_1
3 2020-11-17 12:02:02 marker_1
4 2020-11-17 12:03:03 marker_1
5 2020-11-17 12:04:04 NA      
6 2020-11-17 12:05:05 NA

组合指示标记开始和结束的 POSIXct 列

Combine POSIXct columns indicating start and end of a marker

datetime

r

date

lubridate