[datatable-help] unique

[Matthew Dowle]

Fixed, thanks.

o    unique.data.table now calls the same internal code
     (in C) that grouping calls. This fixes a bug with
     NA in keys when unique is called directly and the key
     contains NA (which may be quite rare). Thanks to
     Damian Betebenner for bug report. unique should also
     now be faster.


On Sat, 2011-06-25 at 19:34 -0500, Damian Betebenner wrote:
> Thanks Matthew
> 
> I was just trying something and saw this.
> 
> I can get what I want using unique with a data.frame. It gives the result one would expect.
> 
> This did make me somewhat scared that using a "by" statement with a data.table on multiple key might skip some unique combinations.
> 
> Damian 
> 
> 
> 
> Damian Betebenner
> Center for Assessment
> PO Box 351
> Dover, NH   03821-0351
>  
> Phone (office): (603) 516-7900
> Phone (cell): (857) 234-2474
> Fax: (603) 516-7910
> 
> dbetebenner at nciea.org
> www.nciea.org
> 
> 
> 
> 
> -----Original Message-----
> From: Matthew Dowle [mailto:mdowlenoreply at virginmedia.com] On Behalf Of Matthew Dowle
> Sent: Saturday, June 25, 2011 8:27 PM
> To: Damian Betebenner
> Cc: datatable-help at lists.r-forge.r-project.org
> Subject: Re: [datatable-help] unique
> 
> The NA in the key seems to be the issue. duplicated.data.table calls
> diff on the columns and assumes no NA in keys.
> Can't remember - did we decide to disallow NA in keys? There were other
> issues with NA in keys and some threads in the past.
> Do you need the key to contain NA?
> Matthew
> 
> 
> On Sat, 2011-06-25 at 16:23 -0500, Damian Betebenner wrote:
> > All,
> > 
> >  
> > 
> > I have a simple question regarding the behavior of unique with more
> > than 1 key. Consider the example:
> > 
> >  
> > 
> > > dt <- data.table(X=c(NA,1,2,3), Y=c(NA,2,1,3))
> > 
> > > dt
> > 
> >       X  Y
> > 
> > [1,] NA NA
> > 
> > [2,]  1  2
> > 
> > [3,]  2  1
> > 
> > [4,]  3  3
> > 
> > > key(dt) <- c("X", "Y")
> > 
> > > unique(dt)
> > 
> >       X  Y
> > 
> > [1,] NA NA
> > 
> > [2,]  2  1
> > 
> > [3,]  3  3
> > 
> >  
> > 
> > If I understand this correctly, unique see rows 2 and 3 of dt as the
> > same. 
> > 
> >  
> > 
> > Is this the behavior one wants?
> > 
> >  
> > 
> > Thanks for any clarification.
> > 
> >  
> > 
> > Damian
> > 
> >  
> > 
> >  
> > 
> > 
> > _______________________________________________
> > datatable-help mailing list
> > datatable-help at lists.r-forge.r-project.org
> > https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/datatable-help
> 
>