tree construction crashes (PR#1698)

Prof Brian D Ripley
Wed, 26 Jun 2002 13:35:52 +0100 (BST)

!) the construction does not `crash', it terminates with an error message,
and you could have used the standard R debugging tools to investigate. Your
incorrect report wasted my quite a lot of my time setting up a debugging
environment to catch a crash.

2) The problem is that you are asking for an exact fit to a large dataset,
and you have integer overflow on your 32-bit (I presume) machine.
What you are doing is not within the design specifcation of tree (nor
rpart). From the help page:

     To produce a tree that fits the data perfectly, set `mindev = 0'
     and `minsize = 2', if the limit on tree depth allows such a tree.

so this is even something discussed on the help page.

3) Reports on contributed packages should be sent or at least copied to
the maintainers, who do not in general have access to R-bugs nor read
R-devel.  In any case, it would be simple good manners.

On Thu, 20 Jun 2002 wrote:

> Full_Name: Petr Savicky
> Version: 1.5.0
> OS: SuSE 7.2
> Submission from: (NULL) (
> The function tree() from the tree package crashes for certain input data.
> All the information needed to reproduce the error (data, script, expected
> error message) may be found at
> Exactly the same error appeared at another computer running some RedHat
> and R-1.4.1.
> -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
> r-devel mailing list -- Read
> Send "info", "help", or "[un]subscribe"
> (in the "body", not the subject !)  To:
> _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Brian D. Ripley,        
Professor of Applied Statistics,
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

r-devel mailing list -- Read
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: