I'm looking for a simple and correct way to change the date/time (POSIXct) format into a time that starts at 00:00:00.

I couldn't find an answer to this in R language, but if I overlooked one, please tell me :)

So I have this :

And I would like this :

Can somebody help me with this?

Thanks :)

CodePudding user response：

You can calculate the difference between the two datetimes in seconds, and add i to a random date starting at "00:00:00", before formatting it to only including the time. See the time column in the reprex underneath:

library(dplyr)
ibrary(lubridate)

df %>% 
  mutate(
    date = lubridate::ymd_hms(date),
    seconds = as.numeric(date - first(date)),
    time = format(
      lubridate::ymd_hms("2022-01-01 00:00:00")   seconds, 
      format = "%H:%M:%S"
    )
  )
#> # A tibble: 2 × 4
#>   date                   v1 seconds time    
#>   <dttm>              <dbl>   <dbl> <chr>   
#> 1 2022-02-16 15:07:15 38937       0 00:00:00
#> 2 2022-02-16 15:07:17 39350       2 00:00:02

^{Created on 2022-03-30 by the reprex package (v2.0.1)}

Note that this will be misleading if you ever have over 24 hours between two datetimes. In these cases you should probably include the date.

Data

df <- tibble::tribble(
  ~date,    ~v1,
  "2022-02-16 15:07:15",    38937,
  "2022-02-16 15:07:17",    39350
)

CodePudding user response：

You can deduct all date/time with the first record of date/time, and change the result to type of time by the hms() function in the hms package.

library(dplyr)
library(hms)

df %>% 
  mutate(`date/time` = hms::hms(as.numeric(as.POSIXct(`date/time`) - as.POSIXct(first(`date/time`)))))

 date/time    v1
1  00:00:00 38937
2  00:00:02 39350

Note that in this method, even if the time difference is greater than 1 day, it'll be reflected in the result, for example:

df <- read.table(header = T, check.names = F, sep = "\t", text = "
date/time   v1
2022-02-16 15:07:15 38937
2022-02-18 15:07:17 39350")

df %>% 
  mutate(`date/time` = hms::hms(as.numeric(as.POSIXct(`date/time`) - as.POSIXct(first(`date/time`)))))

  date/time    v1
1  00:00:00 38937
2  48:00:02 39350