CODEWITHSUNDEEP
rlubridate
A. Hada
A. Hada

Reputation: 105

Converting date variable as character to numeric in r

I have a dataframe with a time period column that is saved as characters and is in the following format:

x
[1] "00:17:31.199"          "1 day 01:37:46.22"     "00:43:11.51"           "01:18:37.721" ...

I would like to convert the values of this column to hours (and hour decimals) 31/ so that the returning column is

[1] 0.28 25.61 0.71 1.3 ...

in which 0.28 is 17/60 hours and 25.61 is equal to 24+1+37/60. Note that everything is saved as characters. I couldn't find any command in baseR or lubridate to figure this out. Any help?

Upvotes: 0

Views: 760

Answers (1)

A. Hada
A. Hada

Reputation: 105

So, I ended up doing the following based on @Mako212 and @thelatemail:

days <- gsub("(\\d+)(?=\\sday).*", "\\1", x, perl=TRUE) # extracting the number of days
hours <- gsub(".*(\\S*\\s+\\S+)", "\\1", x, perl=TRUE) # getting rid of days
ind <- grep("(\\d+)(?=\\sday).*", x, perl=TRUE) # getting the indices of elements that have days

Then I used as.difftime in order to calculate the time difference in hours.

time.hours <- as.difftime(x, format="%H:%M:%OS", units="hours")

Now, I need to calculate the days that I extracted in vector days in terms of hours and add it to time.hours.

for (i in ind) {
time.hours[ind] <- as.numeric(days[ind])*24 +
    as.difftime(hours[ind], format="%H:%M:%OS", units="hours")
}

Note that there might be a more efficient way of doing the above loop but that's how I did it.

Upvotes: 1

Related Questions