Trying to combine dates and times - r

Question

I am trying to combine dates and times. These are from a file when imported, looks like this:
library(tidyverse)
library(lubridate)
bookings <- structure(list(booking_date = structure(c(1549670400, 1550275200,
1550880000, 1551484800, 1552089600, 1552694400), class = c("POSIXct",
"POSIXt"), tzone = "UTC"), start_time = structure(c(-2209043700,
-2209043700, -2209043700, -2209043700, -2209043700, -2209043700
), class = c("POSIXct", "POSIXt"), tzone = "UTC")), row.names = c(NA,
-6L), class = c("tbl_df", "tbl", "data.frame"))
Which looks like this:
# A tibble: 6 x 2
booking_date start_time
<dttm> <dttm>
1 2019-02-09 00:00:00 1899-12-31 08:45:00
2 2019-02-16 00:00:00 1899-12-31 08:45:00
3 2019-02-23 00:00:00 1899-12-31 08:45:00
4 2019-03-02 00:00:00 1899-12-31 08:45:00
5 2019-03-09 00:00:00 1899-12-31 08:45:00
6 2019-03-16 00:00:00 1899-12-31 08:45:00
Obviously the date in the start_time column is wrong. It should be combined with the booking date, so that the first row should read 2019-02-09 08:45:00.
What would the best way of doing that be? I have tried this (based on this other answer), which doesn't really work in my situation.
bookings %>%
select(booking_date, start_time) %>%
mutate(time_2 = as.character(start_time)) %>%
mutate(time_3 = str_sub(time_2, -8, -1)) %>%
mutate(booking_start = dmy(paste(booking_date, time_3)))
Thanks.

Answer 1

If you want to get date for start_time from booking_date a base R approach would be to paste "Date" part from booking_date and "time" part from start_time and convert them to POSIXct.
bookings$start_time <- as.POSIXct(paste(as.Date(bookings$booking_date),
format(bookings$start_time, "%T")))
bookings
# A tibble: 6 x 2
# booking_date start_time
# <dttm> <dttm>
#1 2019-02-09 00:00:00 2019-02-09 08:45:00
#2 2019-02-16 00:00:00 2019-02-16 08:45:00
#3 2019-02-23 00:00:00 2019-02-23 08:45:00
#4 2019-03-02 00:00:00 2019-03-02 08:45:00
#5 2019-03-09 00:00:00 2019-03-09 08:45:00
#6 2019-03-16 00:00:00 2019-03-16 08:45:00
If you want to use it in pipes you can do
library(dplyr)
bookings %>%
mutate(start_time = as.POSIXct(paste(as.Date(booking_date),
format(start_time, "%T"))))

Answer 2

We can also do this with lubridate::date.
date() <- lets you set the date component of a date/time object:
# Set the date component of start_time to be the date component of booking_date
date(bookings$start_time) <- bookings$booking_date
bookings
# A tibble: 6 x 2
booking_date start_time
<dttm> <dttm>
1 2019-02-09 00:00:00 2019-02-09 08:45:00
2 2019-02-16 00:00:00 2019-02-16 08:45:00
3 2019-02-23 00:00:00 2019-02-23 08:45:00
4 2019-03-02 00:00:00 2019-03-02 08:45:00
5 2019-03-09 00:00:00 2019-03-09 08:45:00
6 2019-03-16 00:00:00 2019-03-16 08:45:00
Since it uses assignment (<-), you can't use this first method as part of a pipe. What does work in a pipe, is the update.POSIXt method (see ?DateTimeUpdate), which lets you update the date components of a date, though you have to specify each part of the components specifically:
library(lubridate)
bookings %>%
mutate(date_time = update(start_time,
year = year(booking_date),
month = month(booking_date),
day = day(booking_date)))
booking_date start_time date_time
<dttm> <dttm> <dttm>
1 2019-02-09 00:00:00 1899-12-31 08:45:00 2019-02-09 08:45:00
2 2019-02-16 00:00:00 1899-12-31 08:45:00 2019-02-16 08:45:00
3 2019-02-23 00:00:00 1899-12-31 08:45:00 2019-02-23 08:45:00
4 2019-03-02 00:00:00 1899-12-31 08:45:00 2019-03-02 08:45:00
5 2019-03-09 00:00:00 1899-12-31 08:45:00 2019-03-09 08:45:00
6 2019-03-16 00:00:00 1899-12-31 08:45:00 2019-03-16 08:45:00