[R] stack overflow using step()

Prof Brian Ripley ripley at stats.ox.ac.uk
Mon Oct 17 09:23:51 CEST 2005


Please note the message is about the protect stack, and not the `stack'. 
The message you quote did not come from a current version of R (and you 
have not told us what version you are using, as asked in the posting 
guide).

On Mon, 17 Oct 2005 stich at uni-hohenheim.de wrote:

> Hi,
> I'm using step() for forward regression with 680 main effects and the
> correspomding 230860 interactions.

Hmm.  How many data points?  step() only tries 1000 steps by default.

> However, the calculation is stopped with the message:
> Error: protect(): stack overflow
>
> How can I specify the maximum size of the pointer protection stack when running
> R not from command line but in an emacs-subwindow?

Ask ESS questions on the ESS list please, as the FAQ does ask you.

> How can I perform these calculations when the maximum value accepted is 
> 100000?

which is a lot larger than the default value (which depends on your 
unstated R version).

Without a lot more information it is unclear why the protection stack is 
being used up.  Please show us the last few lines of the output.

-- 
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




More information about the R-help mailing list