round_any on Date objects

[Ivan Alves]

Hi,

I would like to group observations for a boxplot with round_any(), but
it turns out that no method is defined for Date objects (I would like
to group daily observations in weekly of monthly bins). Any
suggestion on which functions/packages I could use? Many thanks in
advance.

Kind regards,
Ivan

[hadley wickham]

Hi Ivan,

Have a look at ?round.Date

Hadley

--
http://had.co.nz/

[Ivan Alves]

Thanks Hadley. Unfortunately "methods for class "Date" are of little
use except to remove fractional days", and therefore does not work for
grouping daily data in weeks or months. Should I tinker around with
as.numeric() and round_any()?

[hadley wickham]

Ooooooh, I had forgotten about that. You might want to have a look at
the code for floor_date and ceiling_date in ggplot2 - I wrote these
methods to work around some of my frustrations with date time
functionality in R. I didn't implement a round function, but I think
you should be able to

--
http://had.co.nz/

[Wayne Folta]

>Thanks Hadley. Unfortunately "methods for class "Date" are of little
>use except to remove fractional days", and therefore does not work for
>grouping daily data in weeks or months. Should I tinker around with
>as.numeric() and round_any()?

If I understand your original question, cut.POSIXt sounds like it could bin your time data for you.

Wayne

[hadley wickham]

Sorry - hit sent too soon.

You should be able to implement round_date by modifying ceiling_date.
You can use the functions like:

floor_date(mydate, "5 day")
floor_date(mydate, "2 week")
floor_date(mydate, "month")

etc.

But I should make round_any generic - I'll add that to my to do.

Hadley

--
http://had.co.nz/

[hadley wickham]

Good point - that's what floor_date and ceiling_date are based on.
Unfortunately there are a few annoying bugs in cut.POSIXt that you
need to work around. I've reported them on R-help, but there doesn't
seem to be any desire to fix them.

Hadley

--
http://had.co.nz/

[Ivan Alves]

In fact (many thanks Wayne!) cut() does the job.

Many thanks to Hadley and Wayne.