R devel - Jun 2003 - Private: Problem with tapply/lapply and sample (PR#3286)

Full_Name: Peter Gedeck
Version: R1.6.2 and R1.7.0
OS: Windows XP
Submission from: (NULL) (194.191.169.72)


Hello,

I marked the bug report Private, as I don't want my email address on the web
server. The problem that I found is best explained using an example.

index <- 1:6
cluster <- c(1,1,1,2,2,3)
tapply(index,cluster,sample)

gives 

$"1"
[1] 2 1 3
$"2"
[1] 4 5
$"3"
[1] 3 2 4 1 6 5

The result for 1 and 2 is correct, the last line should however be 
$"3"
[1] 6
and not return a sample containing all data. Other functions seem to work
correctly,
> tapply(index,cluster,min)
1 2 3 
1 4 6 
> tapply(index,cluster,max)
1 2 3 
3 5 6 
>  tapply(index,cluster,sum)
1 2 3 
6 9 6 

Another example which maybe gives an indication where the problem lies is 
> index <- 1:3
> cluster <- c(1,2,3)
> tapply(index,cluster,sample)
$"1"
[1] 1
$"2"
[1] 2 1
$"3"
[1] 1 3 2

The results should be 1, 2 and 3.

I tried to identify where the error occurs, but I think the error happens in
.Internal(lapply(X,FUN)) in lapply.

Regards,

Peter

peter.gedeck@pharma.novartis.com writes:
> I marked the bug report Private, as I don't want my email address on
the web
> server. 
That doesn't actually work. There used to be a checkbox for it but it
was removed a while ago. I see I forgot to remove the text that
explained how it worked (or didn't work: generally very few people got
to see those reports and that's why the feature got removed.) 

I'll fix the web page shortly.

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk)             FAX: (+45) 35327907

You have already broadcast your address, and it is not reasonable to 
expect help anonymously.

In this case, the bug is in your reading of the help page for sample, 
quite possibly because you didn't read it.

Perhaps this might help your comprehension:

mysample <- function(x,...) {
    print(x)
    sample(x, ...)
}
> tapply(index,cluster,mysample)
[1] 1 2 3
[1] 4 5
[1] 6
$"1"
[1] 1 3 2

$"2"
[1] 5 4

$"3"
[1] 1 3 6 2 5 4
> sample(6)
[1] 5 2 3 4 1 6

Now look up what sample(6) does.

On Wed, 18 Jun 2003 peter.gedeck@pharma.novartis.com wrote:
> Full_Name: Peter Gedeck
> Version: R1.6.2 and R1.7.0
> OS: Windows XP
> Submission from: (NULL) (194.191.169.72)
> 
> 
> Hello,
> 
> I marked the bug report Private, as I don't want my email address on
the web
> server. The problem that I found is best explained using an example.
> 
> index <- 1:6
> cluster <- c(1,1,1,2,2,3)
> tapply(index,cluster,sample)
> 
> gives 
> 
> $"1"
> [1] 2 1 3
> $"2"
> [1] 4 5
> $"3"
> [1] 3 2 4 1 6 5
> 
> The result for 1 and 2 is correct, the last line should however be 
> $"3"
> [1] 6
> and not return a sample containing all data. 
That is exactly what you asked for.  Working as documented is not a bug.
> Other functions seem to work
> correctly,
> 
> > tapply(index,cluster,min)
> 1 2 3 
> 1 4 6 
> > tapply(index,cluster,max)
> 1 2 3 
> 3 5 6 
> >  tapply(index,cluster,sum)
> 1 2 3 
> 6 9 6 
> 
> Another example which maybe gives an indication where the problem lies is 
The finger is still pointing at you!

-- 
Brian D. Ripley,                  ripley@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595