CODEWITHSUNDEEP
rdatetime
Big Russ
Big Russ

Reputation: 1005

Convert factor of format Hh Mm Ss to time duration

I'm working with a data frame that has a variable called "Duration" with values in the format:

1h 7m 46s

0h 16m 41s

..and so on. The column is formatted as a "factor" by default, and I'm wondering how to convert it to an actual duration. I'd like to be able to compute averages and sums of the durations.

Upvotes: 3

Views: 4045

Answers (2)

Jthorpe
Jthorpe

Reputation: 10204

You can extract the hours, minutes, and seconds as follows:

x <- c('1h 7m 46s','0h 16m 41s')
hours <- as.numeric(gsub('^(?:.* )?([0-9]+)h.*$','\\1',x))
minutes <- as.numeric(gsub('^.* ([0-9]+)m.*$','\\1',x))
seconds <- as.numeric(gsub('^.* ([0-9]+)s.*$','\\1',x))
duration_seconds <- seconds + 60*minutes + 60*60*hours

the pattern for minutes is translated as: Starts with (^) zero or more(*) characters(.), followed by a space (), followed by 1 or more(+) digits([0-9]) followed by the letter m (m) followed by zero or more(*) characters(.) to the end of the string($)

Bonus: the (?:.* )? in the regex for hours is a non-capturing group((?: )), which consumes zero or more (*) chartacters (.) followed by a space (). Note that because (?:.* )? is a non-capturing gruop. \\1 still referrs to the number string.

Upvotes: 5

Mike Satteson
Mike Satteson

Reputation: 736

Consider converting the times to strings with an as.character() cast. Once your times are strings you can convert them into DateTime objects with the strptime command, e.g.

> s <- "1h 7m 46s"
> tfmt <- "%Hh %Mm %Ss"
> t1 <- strptime(s, format=tfmt)
> s <- "0h 16m 41s"
> t2 <- strptime(s, format=tfmt)

Having the data in this format is handy as there are then tools for working with them

> t1
[1] "2015-01-30 01:07:46"
> t2
[1] "2015-01-30 00:16:41"
> t1 - t2
Time difference of 51.08333 mins
> difftime(t1, t2, units="secs")
Time difference of 3065 secs

Upvotes: 4

Related Questions