用非标准格式格式化日期

Question

I'm relatively new to this site so forgive me if my question is a bit vague for you guys. I also realize there are many threads on this topic, yet I feel they do not answer my question specifically since they are almost all about changing yy/mm/dd to dd/mm/yy or vice versa.

In short what do i want? I want my current format changed into only year.

I have a column full of dates of this format.

31OCT2016:23:52:00.000

I've seen in many topics you can use format commands but they go something like this;

dates <- c("05/27/84", "07/07/05")

I have over 100.000 observations so this can't be done manually. So I tried;

mydata$dates <- format(as.Date(mydata$dates), "%Y")

But that didn't work. I saw on this website the proper values

http://www.statmethods.net/input/dates.html

But it did not say anything on how to get rid of hours minutes and seconds.

So what is the easiest way to strip it all down to year only?

Answer 1

Lubridate is your friend. To be precise, the function dmy_hms :

I'll generate some sample data which has the same format as your example so my code is reproducible. Don't worry about it too much. For your purposes, you can jump right to the conversion part.

#------------------------------------------------------------------------------------------
#This code block is entirely for generating reproducible sample data

d <- sample(1:27,10,T)
mon <- toupper(sample(month.abb,10,T))
y <- sample(2000:2017,10,T)
h <- sample(0:23,10,T)
min <- sample(0:59,10,T)
s <- sample(0:59,10,T)

#load package
library(lubridate)

dts <- sprintf('%02d%s%s:%s:%s:%s.000',d,mon,y,h,min,s)

> dts
 [1] "01JAN2012:12:6:53.000"  "01NOV2010:0:19:47.000"  "03SEP2000:9:45:3.000"   "25NOV2009:21:39:57.000" "08DEC2015:19:27:36.000"
 [6] "23MAR2009:13:39:40.000" "03JUN2010:14:54:50.000" "03APR2002:6:34:45.000"  "19NOV2012:5:17:29.000"  "02FEB2003:0:3:59.000" 

#------------------------------------------------------------------------------------------

So basically the variable dts is your column full of dates which you want to convert:

#conversion

> dmy_hms(dts)
 [1] "2012-01-01 12:06:53 UTC" "2010-11-01 00:19:47 UTC" "2000-09-03 09:45:03 UTC" "2009-11-25 21:39:57 UTC"
 [5] "2015-12-08 19:27:36 UTC" "2009-03-23 13:39:40 UTC" "2010-06-03 14:54:50 UTC" "2002-04-03 06:34:45 UTC"
 [9] "2012-11-19 05:17:29 UTC" "2003-02-02 00:03:59 UTC"

And then to get just the years, you can use the year function:

> year(dmy_hms(dts))
 [1] 2012 2010 2000 2009 2015 2009 2010 2002 2012 2003

---

So assuming you want to do everything inside the data.frame , your code could look like this:

# example dataframe
dframe <- data.frame(variable=c('A','B','C'),dates=sample(dts,3))

This is a data frame with some variable and the column with the dates.

> dframe
  variable                 dates
1        A  15JAN2000:0:37:6.000
2        B 13DEC2016:8:34:28.000
3        C 18AUG2005:2:27:16.000

So to convert the dates, we can simply do dframe$dates <- year(dmy_hms(dframe$dates))

If we look at dframe again, we can see that the conversion was successful:

> dframe
  variable dates
1        A  2000
2        B  2016
3        C  2005