在 R 中添加白天/黄昏/夜晚/黎明的信息以跟踪数据答案

【问题标题】：Adding information on day/dusk/night/dawn for tracking data in R在 R 中添加白天/黄昏/夜晚/黎明的信息以跟踪数据
【发布时间】：2021-05-31 08:24:17
【问题描述】：

我有一个数据集，其中包含几个月的日期时间、纬度和经度变量的跟踪数据，如下所示：

> start <- as.POSIXct("2018-08-01 00:00:00", format="%Y-%m-%d %H:%M:%S", tz="UTC")
> datetime <- seq(from = start, length.out = 2880, by = "5 mins")
> lat<-rep(seq(from=50, to= 30, length.out = 10), each=288)
> lon<-rep(seq(from=110, to= 70, length.out = 10), each=288)
> data<-cbind.data.frame(datetime, lat, lon)
> head(data)
             datetime lat lon
1 2018-08-01 00:00:00  50 110
2 2018-08-01 00:05:00  50 110
3 2018-08-01 00:10:00  50 110
4 2018-08-01 00:15:00  50 110
5 2018-08-01 00:20:00  50 110
6 2018-08-01 00:25:00  50 110

我想根据来自getSunlightTimes 的信息添加一个新列，其中包含有关一天中的时间段的信息，即白天、黄昏、夜晚或黎明：

> data$date<-as.Date(data$datetime)
> sun<-getSunlightTimes(data=data, tz="UTC", keep=c("sunrise","sunset","night","nightEnd"))
> head(sun)
        date lat lon             sunrise              sunset               night
1 2018-08-01  50 110 2018-07-31 21:09:50 2018-08-01 12:25:18 2018-08-01 15:01:56
2 2018-08-01  50 110 2018-07-31 21:09:50 2018-08-01 12:25:18 2018-08-01 15:01:56
3 2018-08-01  50 110 2018-07-31 21:09:50 2018-08-01 12:25:18 2018-08-01 15:01:56
4 2018-08-01  50 110 2018-07-31 21:09:50 2018-08-01 12:25:18 2018-08-01 15:01:56
5 2018-08-01  50 110 2018-07-31 21:09:50 2018-08-01 12:25:18 2018-08-01 15:01:56
6 2018-08-01  50 110 2018-07-31 21:09:50 2018-08-01 12:25:18 2018-08-01 15:01:56
             nightEnd
1 2018-07-31 18:33:13
2 2018-07-31 18:33:13
3 2018-07-31 18:33:13
4 2018-07-31 18:33:13
5 2018-07-31 18:33:13
6 2018-07-31 18:33:13

所以黄昏对应于sunset和night之间的datetime值，nightEnd和sunrise之间的黎明，sunrise和sunset之间的白天和night和nightEnd之间的夜晚.

我试过了：

> data$period<-rep(" ", length.out=nrow(data))
> data$period[which(data$datetime>sun$sunrise & data$datetime<sun$sunset)]<-"day"
> data$period[which(data$datetime>sun$sunset & data$datetime<sun$night)]<-"dusk"
> data$period[which(data$datetime>sun$nightEnd & data$datetime<sun$sunrise)]<-"dawn"
> data$period[which(data$period==" ")]<-"night"

但这会在几天之间的转换中引起问题。有人有什么建议吗？

亲切的问候

【问题讨论】：

请让 SO 用户知道解决方案是否有效。如果它们不起作用，也许我们可以修复它们。

标签： r date datetime posixct date-sunrise

【解决方案1】：

我要做的是进行左连接，即将日落/日出数据附加到您的实际日期并在那里发挥您的魔力。

library(suncalc)

start <- as.POSIXct("2018-08-01 00:00:00", format = "%Y-%m-%d %H:%M:%S", 
                    tz = "UTC")
datetime <- seq(from = start, length.out = 2880, by = "5 mins")
lat <- rep(seq(from = 50, to = 30, length.out = 10), each = 288)
lon <- rep(seq(from = 110, to = 70, length.out = 10), each = 288)
data <- cbind.data.frame(datetime, lat, lon)

data$date <- as.Date(data$datetime)
sun <- getSunlightTimes(data = data, tz = "UTC", 
                        keep = c("sunrise", "sunset", "night", "nightEnd"))

xy <- merge(x = data, 
            y = sun[, c("date", "sunrise", "sunset", "night", "nightEnd")], 
            by = "date")

xy$period <- rep(" ", length.out = nrow(xy))
xy$period[xy$datetime > xy$sunrise & xy$datetime < xy$sunset] <- "day"
xy$period[xy$datetime > xy$sunset & xy$datetime < xy$night] <- "dusk"
xy$period[xy$datetime > xy$nightEnd & xy$datetime < xy$sunrise] <- "dawn"
xy$period[xy$period == " "] <- "night"

set.seed(357)
xy[sample(1:nrow(xy), size = 7), ]

             date            datetime      lat       lon             sunrise              sunset               night            nightEnd period
396824 2018-08-05 2018-08-05 18:45:00 41.11111  92.22222 2018-08-04 22:51:28 2018-08-05 13:05:23 2018-08-05 14:55:08 2018-08-04 21:01:43  night
407953 2018-08-05 2018-08-05 22:00:00 41.11111  92.22222 2018-08-04 22:51:28 2018-08-05 13:05:23 2018-08-05 14:55:08 2018-08-04 21:01:43  night
685732 2018-08-09 2018-08-09 06:25:00 32.22222  74.44444 2018-08-09 00:23:28 2018-08-09 13:54:45 2018-08-09 15:25:18 2018-08-08 22:52:56    day
638810 2018-08-08 2018-08-08 16:50:00 34.44444  78.88889 2018-08-08 00:01:06 2018-08-08 13:41:48 2018-08-08 15:16:02 2018-08-07 22:26:53  night
392649 2018-08-05 2018-08-05 17:35:00 41.11111  92.22222 2018-08-04 22:51:28 2018-08-05 13:05:23 2018-08-05 14:55:08 2018-08-04 21:01:43  night
165451 2018-08-02 2018-08-02 23:50:00 47.77778 105.55556 2018-08-01 21:36:31 2018-08-02 12:34:05 2018-08-02 14:53:05 2018-08-01 19:17:31  night
159367 2018-08-02 2018-08-02 22:05:00 47.77778 105.55556 2018-08-01 21:36:31 2018-08-02 12:34:05 2018-08-02 14:53:05 2018-08-01 19:17:31  night

【讨论】：

【解决方案2】：

一个 tidyverse 选项 - 就像 Roman Luštrik 的解决方案一样 - 使用 left_join。

library(dplyr)
library(lubridate)

df1 %>%
  mutate(date = date(datetime)) %>%
  left_join(df2) %>%
  distinct() %>%
  mutate(period = case_when(datetime %within% interval(sunset, night) ~ 'dusk',
                            datetime %within% interval(nightEnd, sunrise) ~ 'dawn',
                            datetime %within% interval(sunrise, sunset) ~ 'day',
                            datetime %within% interval(night, nightEnd) ~ 'night')) %>%
  select(datetime, lat, lon, period)

# # A tibble: 6 x 4
#   datetime              lat   lon period
#   <dttm>              <dbl> <dbl> <chr> 
# 1 2018-08-01 00:00:00    50   110 day   
# 2 2018-08-01 00:05:00    50   110 day   
# 3 2018-08-01 00:10:00    50   110 day   
# 4 2018-08-01 00:15:00    50   110 day   
# 5 2018-08-01 00:20:00    50   110 day   
# 6 2018-08-01 00:25:00    50   110 day

数据

df1 <- structure(list(datetime = structure(c(1533081600, 1533081900, 
1533082200, 1533082500, 1533082800, 1533083100), class = c("POSIXct", 
"POSIXt"), tzone = "UTC"), lat = c(50, 50, 50, 50, 50, 50), lon = c(110, 
110, 110, 110, 110, 110)), row.names = c(NA, -6L), class = c("tbl_df", 
"tbl", "data.frame"))

df2 <- structure(list(date = structure(c(17744, 17744, 17744, 17744, 
17744, 17744), class = "Date"), lat = c(50, 50, 50, 50, 50, 50
), lon = c(110, 110, 110, 110, 110, 110), sunrise = structure(c(1533071390, 
1533071390, 1533071390, 1533071390, 1533071390, 1533071390), class = c("POSIXct", 
"POSIXt"), tzone = "UTC"), sunset = structure(c(1533126318, 1533126318, 
1533126318, 1533126318, 1533126318, 1533126318), class = c("POSIXct", 
"POSIXt"), tzone = "UTC"), night = structure(c(1533135716, 1533135716, 
1533135716, 1533135716, 1533135716, 1533135716), class = c("POSIXct", 
"POSIXt"), tzone = "UTC"), nightEnd = structure(c(1533061993, 
1533061993, 1533061993, 1533061993, 1533061993, 1533061993), class = c("POSIXct", 
"POSIXt"), tzone = "UTC")), row.names = c(NA, -6L), class = c("tbl_df", 
"tbl", "data.frame"))

【讨论】：