Message-ID: <x2odzow4ap.fsf@viggo.kubism.ku.dk>
Date: 2006-03-30T11:38:38Z
From: Peter Dalgaard
Subject: function min does not return correct result if .Machine$integer.max is involved (PR#8731)
In-Reply-To: <20060330103417.A3A2E19066@slim.kubism.ku.dk>
schlather at hsu-hh.de writes:
> Full_Name: Martin Schlather
> Version: 2.2.0 and alpha 2.3.0 (06/3/29)
> OS: Linux (x86_64 and Intel)
> Submission from: (NULL) (139.11.183.106)
>
>
> > min(.Machine$integer.max, 10^20)
> [1] 1e+20
>
> > min(as.integer(.Machine$integer.max), 10^20)
> [1] 1e+20
>
>
> but
> > min(.Machine$integer.max + 0, 10^20)
> [1] 2147483647
>
> > min(as.integer(.Machine$integer.max - 1), 10^20)
> [1] 2147483646
>
> > min(as.double(.Machine$integer.max), 10^20)
> [1] 2147483647
I have a vague recollection that we might have used
.Machine$integer.max to represent "integer infinity" at some point. If
so, then the results make some sense, but we don't seem to have
similar conventions in any other places that I can think of (i.e.
as.integer(Inf) is NA, etc.)
--
O__ ---- Peter Dalgaard ?ster Farimagsgade 5, Entr.B
c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K
(*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk) FAX: (+45) 35327907