rdatetimedplyrlubridatezoo

R Lubridate: Round/Snap datetime to the nearest arbitrary time of day?

I have a list of datetimes like the following:

data <- data.frame(datetimes = c(ymd_hms("2024-01-01 04:12:35"), 
                                 ymd_hms("2024-04-01 14:52:20"), 
                                 ymd_hms("2024-03-18 23:43:56")))

I would like to round/snap each of these date times to the next occurrence of a specified time of day.

For example, if I choose a time of 10:30:00 I would like to round up each of these dates so that:

2024-01-01 04:12:35 becomes -> 2024-01-01 10:30:00.

2024-04-01 14:52:20 becomes -> 2024-04-02 10:30:00.

2024-03-18 23:43:56 becomes -> 2024-03-19 10:30:00.

Is there a way to easily do this?

Thanks

Solution

  • 1) Add 1 to the date if the datetime is past 10:30:00 and paste 10:30:00 and convert back to POSIXct.

    library(dplyr)
library(lubridate)

cutoff <- "10:30:00" # specify as string
data %>%
  mutate(datetimes2 = 
    (as.Date(datetimes) + (sub(".* ", "", datetimes) > cutoff)) %>%
    paste(cutoff) %>% 
    ymd_hms)

##             datetimes          datetimes2
## 1 2024-01-01 04:12:35 2024-01-01 10:30:00
## 2 2024-04-01 14:52:20 2024-04-02 10:30:00
## 3 2024-03-18 23:43:56 2024-03-19 10:30:00

    2) Alternately subtract 10.5*3600, extract the date, add 1, convert back to POSIXct and add the seconds back on.

    cutoff <- 10.5 * 3600 # specify in seconds
data %>%
  mutate(datatimes2 = as.POSIXct(as.Date(datetimes-cutoff)+1) + cutoff)

##             datetimes          datatimes2
## 1 2024-01-01 04:12:35 2024-01-01 10:30:00
## 2 2024-04-01 14:52:20 2024-04-02 10:30:00
## 3 2024-03-18 23:43:56 2024-03-19 10:30:00