Convert timestamp by milliseconds

[user124578]

I would like to convert timestamp by milliseconds. Can i still do this with the lubridate package?

library(lubridate)
library(tidyr)
library(dplyr)

df <- data.frame(timestamp = c("1 2018-11-08T07:41:55.921Z",
                               "2 2018-11-08T07:42:29.842Z",
                               "3 2018-11-08T07:42:29.845Z",
                               "4 2018-11-08T07:42:29.845Z",
                               "5 2018-11-08T07:43:13.957Z"))

df %>%
  
  #note your timestamp has index info, separating that out
  separate(timestamp, c('index', 'datetime'), sep = ' ') %>% 
  
  # converting the form of "2018-11-08T07:41:55.921Z" to POSIXct date-time objects
  # check out `?ymd_hms` docs if you want to force a timezone other than UTC
  mutate(time_clean = lubridate::ymd_hms(datetime)) %>% 
  
  mutate(
    # see https://stat.ethz.ch/R-manual/R-devel/library/base/html/strptime.html 
    # formating details
    date = format(time_clean,format = '%F'),
    time = format(time_clean,format = '%T')
  )

[andresrcs]

You are not using lubridate for printing the timestamp to any particular format, that is done by base R format() function, and yes, you can get miliseconds too. Just have in mind that the result is a character vector and you are not going to be able to directly perform calculations with it or use it on an axis with a continuous scale.

library(dplyr)
library(tidyr)
library(lubridate)

df <- data.frame(timestamp = c("1 2018-11-08T07:41:55.921Z",
                               "2 2018-11-08T07:42:29.842Z",
                               "3 2018-11-08T07:42:29.845Z",
                               "4 2018-11-08T07:42:29.845Z",
                               "5 2018-11-08T07:43:13.957Z"))
options(digits.secs=3)
df %>% 
    separate(timestamp, c('index', 'datetime'), sep = ' ') %>%
    mutate(time_clean = lubridate::ymd_hms(datetime)) %>% 
    mutate(date = format(time_clean,format = '%F'), 
           time = format(time_clean,format = '%H:%M:%OS')) %>%
    select(date, time)
#>         date         time
#> 1 2018-11-08 07:41:55.921
#> 2 2018-11-08 07:42:29.842
#> 3 2018-11-08 07:42:29.845
#> 4 2018-11-08 07:42:29.845
#> 5 2018-11-08 07:43:13.957

^{Created on 2019-01-26 by the reprex package (v0.2.1)}

[user124578]

Many Thanks. Can I convert it back to time format to perform any calculation? The reason I want to change it to show in milliseconds so, I can merge the data with the timestamp to another data frame. Merging to the nearest minutes, I missing a lot of values.

[user124578]

I am getting this message when i try to convert the whole dataframe:

Warning message:
Expected 2 pieces. Missing pieces filled with NA in 1543681 rows [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, ...]

[andresrcs]

Can I convert it back to time format to perform any calculation?

No need for convert it back, the POSIXct object is already stored with full precision, you are just not seen it because of your digits.secs option.

library(dplyr)
library(tidyr)
library(lubridate)

df <- data.frame(timestamp = c("1 2018-11-08T07:41:55.921Z",
                               "2 2018-11-08T07:42:29.842Z",
                               "3 2018-11-08T07:42:29.845Z",
                               "4 2018-11-08T07:42:29.845Z",
                               "5 2018-11-08T07:43:13.957Z"))
options(digits.secs=3) # Change digits.sec to show milliseconds
df %>% 
    separate(timestamp, c('index', 'datetime'), sep = ' ') %>%
    mutate(time_clean = lubridate::ymd_hms(datetime))
#>   index                 datetime              time_clean
#> 1     1 2018-11-08T07:41:55.921Z 2018-11-08 07:41:55.921
#> 2     2 2018-11-08T07:42:29.842Z 2018-11-08 07:42:29.842
#> 3     3 2018-11-08T07:42:29.845Z 2018-11-08 07:42:29.845
#> 4     4 2018-11-08T07:42:29.845Z 2018-11-08 07:42:29.845
#> 5     5 2018-11-08T07:43:13.957Z 2018-11-08 07:43:13.957

About your warning message, I can't tell why is happening because I don't have access to your data, but is very likely that you have some missing values.

[user124578]

Prefect. Thanks Very Much for your help. That's very helpful.