[R] converting character vector "hh:mm" to chron or strptime 24 clock time vectors

Alex Anderson complicado79 at yahoo.com.au
Tue Feb 16 11:47:58 CET 2010


Hi All,
I am attempting to work with some data from loggers. I have read in a 
.csv exported from MS Access that already has my dates and times (in 24 
clock format), (with StringsAsFactors=FALSE).

 > head(tdata)

   LogData       date              time
1    77.16     2008/04/24     02:00
2    61.78     2008/04/24     04:00
3    75.44     2008/04/24     06:00
4    89.43     2008/04/24     08:00
5    95.83     2008/04/24     10:00
6    96.88     2008/04/24     24:00

I wish to be able to summarise the data using the character vectors 
$data and $time (daily, monthly averages, maxima of my $LogData for 
example) so I am trying to get R to recognise the $date and $time 
columns as valid dates and times. Using...

 > tdata$date2 = as.Date(as.character(tdata$date))

I can get a new column of valid dates, but neither:

 > tdata$time2= strptime(tdata$time,"%k")

Error in `$<-.data.frame`(`*tmp*`, "time2", value = list(sec = c(0, 0,  :
  replacement has 9 rows, data has 10

nor trying:
 
 >  tdata$time2=chron(times=as.character(tdata$time, format= "hh:mm"))

In addition: Warning messages:
1: In unpaste(times, sep = fmt$sep, fnames = fmt$periods, nfields = 3) :
  wrong number of fields in entry(ies) 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
2: In convert.times(times., fmt) : NAs introduced by coercion
3: In convert.times(times., fmt) : NAs introduced by coercion
4: In convert.times(times., fmt) : NAs introduced by coercion

gives me any valid times from my time vector.  the Chron documentation 
doesn't mention 24 clocks, strptime neither, and the Rnews issue 1/4 
with an article about time is no help... Any thoughts would be much 
appreciated.
regards

Alex Anderson
James Cook University
Townsville, Australia



More information about the R-help mailing list