How to categorise time values in R

Question

How can I categorise time values to month buckets in R?
Now

time                     
2014-01-31 20:46 +0000   
2014-01-31 10:46 +0000   
2014-02-21 20:16 +0000   
2014-02-11 12:36 +0000

my goal

time                     month
2014-01-31 20:46 +0000   Jan
2014-01-31 10:46 +0000   Jan
2014-02-21 20:16 +0000   Feb
2014-02-11 12:36 +0000   Feb

`months(as.POSIXct(time))`. Google seem to be on strike today. — David Arenburg, Oct 11 '15 at 20:10

RHA · Accepted Answer · 2015-10-11T20:29:58.610

1

time <- c("2014-01-31 20:46 +0000",   
          "2014-01-31 10:46 +0000",   
          "2014-02-21 20:16 +0000",   
          "2014-02-11 12:36 +0000")

library(lubridate)
# first convert it to a proper date-time format and put it in a dataframe
df <- as.data.frame(as.POSIXct(time))
#then simply calculate month
df$month <- month(time, label= TRUE)

Which gives:

      as.POSIXct(time)  month
1 2014-01-31 20:46:00   Jan
2 2014-01-31 10:46:00   Jan
3 2014-02-21 20:16:00   Feb
4 2014-02-11 12:36:00   Feb

edited Oct 11 '15 at 20:29

answered Oct 11 '15 at 20:20

RHA

3,677
4
25
48

Your example works. I have an issue with my column. When I use your code example and type `typeof(time)` that's what I get back `[1] "character"`. When I try it with my data set I get this message `[1] "closure"`. Is it because I read a CSV file? – co7000 Oct 11 '15 at 21:27
@cosmin Probably. Look at (http://stackoverflow.com/questions/15057202/read-csv-and-colclasses?lq=1) – RHA Oct 12 '15 at 07:58

score 0 · Answer 2 · answered Oct 12 '15 at 12:04

Here's my approach:

data <- read.csv("~/data.csv", 
                header=TRUE, 
                dec=".", 
                sep=",", 
                stringsAsFactors=FALSE)

time <- data$time
time <- as.Date(time, format="%Y-%m-%d %H:%M")
time

data$year <- as.factor(format(time, format ="%Y"))
data$month <- as.factor(format(time, format ="%m"))
data$day <- as.factor(format(time, format ="%d"))

How to categorise time values in R

2 Answers2