[R] cut.POSIXt misconception/feature/bug?

[Brian Diggs]

On 3/10/2010 11:26 PM, Petr PIKAL wrote:
> Hi
> 
> Thanks for clarification. Actually I knew that with first case I get some 
> data with NAs at the beginning and at the end. Maybe my English is not 
> good enough to understand that to get vector of dates split to several 
> chunks I need to put also end date and last date to get the whole vector.
> 
> This is what help page for cut.POSIXt says about breaks and although I 
> read it carefully I did not find any mention that at least 2 values are 
> necessary. I did not connected it with information from cut help page, 
> sorry
> 
> breaks: a vector of cut points _or_ number giving the number of
>         ^^^^^^^^^^^^^^^^^^^^^^
> which can be vector of length one. However thinking about it more 
> thoroughly vector of length one is probably the same as one number with 
> respect of its interpretation. 

Since everything is a vector, even a single number is technically a vector of length 1.  But that is a somewhat subtle point.  I think the description from cut is better: "either a numeric vector of two or more cut points or a single number (greater than or equal to 2) giving the number of intervals into which x is to be cut"

> Therefore I also missed the clue that I need not only
> 
> Details:
> 
>      Using both ‘right = TRUE’ and ‘include.lowest = TRUE’ will
>      include both ends of the range of dates.
> 
> but also
> 
>  br<-dat[c(1, 23, 42,60)]
> 
> To get the whole vector of cut dates without NAs in both ends.
> 
> Maybe
> 
> breaks: a vector of 2 or more cut points _or_ number giving the number 
> of...
> 
> Using both ‘right = TRUE’ and ‘include.lowest = TRUE’ together with 
> starting and ending date will include both ends of the range of dates.
> 
> could make help page more digestable.

I think that the comment in details has more to do with how to include the endpoints of breaks in the included ranges rather than the extremes of the dates (x).  I think a better clarification would be to say "Using both ‘right = TRUE’ and ‘include.lowest = TRUE’ will include both ends of the range of breaks".  That makes it clear that the dates that are referred to there are the breaks, not x.

> Thank you.
> 
> Petr

I've included a patch against cut.POSIXt.Rd with these proposed changes.


--
Brian Diggs, Ph.D.
Senior Research Associate, Department of Surgery, Oregon Health & Science University