R Date time conversion

Question

I am new to R and I have a data frame with date time as variable. For every hour each day temperature is recorded, and date time is in format of YYYY-MM-DD 00:00:00. Now I would like to convert the time into a factor ranging from 0 to 23 each day. So For each day my new column should have factors 0 to 23. Could anyone help me with this? My 2015-01-01 00:00:00, should give me 0, while 2015-01-01 01:00:00, should give me 1 and so on. Also my 2015-01-02 00:00:00 should be 0 again.

Answer 1

You can convert your timestamp into a POSIXlt object. Once you have that, you can obtain the hour directly like this:

> timestamp <- as.POSIXlt("2015-01-01 00:00:00")
> timestamp
[1] "2015-01-01 MYT"
> timestamp$hour
[1] 0

Answer 2

Using a sample data, one way would be the following.

mydf <- data.frame(id = c(1,1,1,2,2,1,1),
                   event = c("start", "valid", "end", "start", "bad", "start", "bad"),
                   time = as.POSIXct(c("2015-05-16 20:46:53", "2015-05-16 20:46:56", "2015-05-16 21:46:59",
                                   "2015-05-16 22:46:53", "2015-05-16 22:47:00", "2015-05-16 22:49:05",
                                   "2015-05-16 23:49:09"), format = "%Y-%m-%d %H:%M:%S"),
                   stringsAsFactors = FALSE)

library(dplyr)
mutate(mydf, group = factor(format(time, "%H")))

#  id event                time group
#1  1 start 2015-05-16 20:46:53    20
#2  1 valid 2015-05-16 20:46:56    20
#3  1   end 2015-05-16 21:46:59    21
#4  2 start 2015-05-16 22:46:53    22
#5  2   bad 2015-05-16 22:47:00    22
#6  1 start 2015-05-16 22:49:05    22
#7  1   bad 2015-05-16 23:49:09    23

Answer 3

Tim's answer using POSIXlt is probably the best option, but here's a regex way just in case:

> times <- c("2015-01-01 00:00:00", "2015-01-01 01:00:00", "2015-01-02 00:00:00")
> regmatches(times, regexpr("(?<=-\\d{2} )\\d{2}", times, perl=TRUE))
[1] "00" "01" "00"

With the extracted hours you can make them factors or integers as necessary.

Answer 4

@Sairam, in addition to @jazzurro's use of 'dplyr' (which, like jazzurro, many R-insitas routinely use)...in the future, if you need/want a simple & powerful way to manipulate dates, you're encouraged to gain familiarity with another package: 'lubridate.'

lubridate makes working with dates a snap. Hope this helps and best regards on your project.