R help - May 2004 - bug in cor (..., use= ...)?

Dear R users,

I have not found anything on this in the archives. Does anyone know whehther
the parameter use= is not functioning in cor or enlighten me what it is
supposed to do?

My R version is "R version 1.8.1, 2003-11-21" on Windows 2000. I am
hoping
to be able to update to 1.9.1 as soon as it has appeared (we are not allowed
here to install software on our own and thus I am trying to be able to have
the .1 versions installed ...).

Test code:

x <- 1:10
y <- 2:11

x [1] <- NA
y [10] <- 12

cor (x, y, use= 'all.obs', method= 'kendall')
cor (x, y, use= 'complete.obs', method= 'kendall')
cor (x, y, use= 'pairwise.complete.obs', method= 'kendall')

As I understand, the first one of this should result in an error which it
does not. All the results are the same and seemingly treat the NA as if it
was 0.

Any ideas are appreciated.

Thanks and regards, Lorenz
- 
Lorenz Gygax, Dr. sc. nat.
Tel: +41 (0)52 368 33 84 / lorenz.gygax at fat.admin.ch      

Tag der offenen T??r, 11./12. Juni 2004: http://www.fat.ch/2004

Center for proper housing of ruminants and pigs
Swiss Veterinary Office
agroscope FAT T??nikon, CH-8356 Ettenhausen / Switzerland
Fax : +41 (0)52 365 11 90 / Tel: +41 (0)52 368 31 31

>From the NEWS file for 1.9.0
    o	The cor() function did not remove missing values in the
	non-Pearson case.

Your example works correctly there.  (I am fairly sure this has been 
discussed on the mailing lists.)

On Mon, 24 May 2004 Lorenz.Gygax at fat.admin.ch wrote:
> Dear R users,
> 
> I have not found anything on this in the archives. Does anyone know
whehther
> the parameter use= is not functioning in cor or enlighten me what it is
> supposed to do?
> 
> My R version is "R version 1.8.1, 2003-11-21" on Windows 2000. I
am hoping
> to be able to update to 1.9.1 as soon as it has appeared (we are not
allowed
> here to install software on our own and thus I am trying to be able to have
> the .1 versions installed ...).
> 
> Test code:
> 
> x <- 1:10
> y <- 2:11
> 
> x [1] <- NA
> y [10] <- 12
> 
> cor (x, y, use= 'all.obs', method= 'kendall')
> cor (x, y, use= 'complete.obs', method= 'kendall')
> cor (x, y, use= 'pairwise.complete.obs', method= 'kendall')
> 
> As I understand, the first one of this should result in an error which it
> does not. All the results are the same and seemingly treat the NA as if it
> was 0.
-- 
Brian D. Ripley,                  ripley at 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