R - converting date and time fields to POSIXct with HHMMSS format

Flyto picture Flyto · Apr 30, 2013 · Viewed 38.2k times · Source

I have a data file which has three columns thus:

20010101 000000  0.833
20010101 000500  0.814
20010101 001000  0.794
20010101 001500  0.772
...

As is fairly clear to human eyes, the first two are date and time. I need to convert them into a POSIXct (or something else if it's better, but my limited past experience of dealing with timestamps in R is to use POSIXct). Normally, having pulled it in with read.table, I would use:

df$DateTime <- as.POSIXct(paste(df$Date, df$Time), format="%Y%m%d %H%M%S")

However, the second column seems to lose its leading zeroes (probably through a type coercion?), and thus it doesn't work correctly.

I've looked at Combine date as integer and time as factor to POSIXct in R and Converting two columns of date and time data to one, but both are using times with delimiters such as :, and so don't have the same problem.

How can I convert these columns to a POSIXct, please?

Answer

Dirk Eddelbuettel picture Dirk Eddelbuettel · Apr 30, 2013

You were very close. The following "simply" forces the first two columns to be read as character strings, which saves the leading zeros.

R> df <- read.table(text="20010101 000000  0.833
20010101 000500  0.814
20010101 001000  0.794
20010101 001500  0.772", 
+ header=FALSE, colClasses=c("character", "character", "numeric"), 
+ col.names=c("Date", "Time", "Val"))
R> df
      Date   Time   Val
1 20010101 000000 0.833
2 20010101 000500 0.814
3 20010101 001000 0.794
4 20010101 001500 0.772

Now what you were attempting "just works":

R> df$DateTime <- as.POSIXct(paste(df$Date, df$Time), format="%Y%m%d %H%M%S")
R> df
      Date   Time   Val            DateTime
1 20010101 000000 0.833 2001-01-01 00:00:00
2 20010101 000500 0.814 2001-01-01 00:05:00
3 20010101 001000 0.794 2001-01-01 00:10:00
4 20010101 001500 0.772 2001-01-01 00:15:00
R>