[Rd] ave returns wrong data type (PR#13664)

Dimitris Rizopoulos d.rizopoulos at erasmusmc.nl
Mon Apr 20 11:53:09 CEST 2009


Note that according to ?ave, the first argument of ave(), 'x' should be 
a *numeric* vector. In your case 'x' is not numeric, it is a character 
vector. So I think that ave() works as documented, i.e., if you supply 
as first argument a numeric vector, then you do get as an output a 
numeric vector.

Best,
Dimitris


brenbarn at brenbarn.net wrote:
> Full_Name: Brendan Barnwell
> Version: 2.9.0
> OS: Windows XP Pro
> Submission from: (NULL) (71.102.131.29)
> 
> 
>    The ave() function returns an incorrect datatype.  Specifically, ave(x, g, f)
> always returns a vector with the same mode as x, rather than using the mode of
> the vector returned by f.  Observe:
> 
>> x
>  [1] "A" "B" "C" "A" "B" "C" "A" "B" "C" "A" "B" "C" "A" "B" "C" "A" "B" "C" "A"
> "B" "C" "A" "B" "C" "A" "B" "C" "A" "B" "C"
>> g
>  [1] "X" "Y" "X" "Y" "X" "Y" "X" "Y" "X" "Y" "X" "Y" "X" "Y" "X" "Y" "X" "Y" "X"
> "Y" "X" "Y" "X" "Y" "X" "Y" "X" "Y" "X" "Y"
>> ave(x, g, FUN=length)
>  [1] "15" "15" "15" "15" "15" "15" "15" "15" "15" "15" "15" "15" "15" "15" "15"
> "15" "15" "15" "15" "15" "15" "15" "15" "15"
> [25] "15" "15" "15" "15" "15" "15"
> 
>    Even though the length() function returns a vector of integers, ave()
> inappropriately converts this to a character vector.  The bug is due to this
> line in the definition of ave(): 
> 
> split(x, g) <- lapply(split(x, g), FUN)
> 
>    By sticking the result of the lapply back into the original argument x, it
> coerces that result to the type of that argument.  This contradicts the
> documentation, which says that the value of ave() is "a numeric vector".  I
> would suggest that this documentation itself doesn't describe the desired
> behavior.  The result vector should be of the type returned by FUN (just as it
> is for tapply).  Otherwise it is impossible to use ave() to compute summary
> statistics whose type differs from that of the argument.
> 
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
> 

-- 
Dimitris Rizopoulos
Assistant Professor
Department of Biostatistics
Erasmus University Medical Center

Address: PO Box 2040, 3000 CA Rotterdam, the Netherlands
Tel: +31/(0)10/7043478
Fax: +31/(0)10/7043014



More information about the R-devel mailing list