buglet in lm.influence() -- easy fix but... (PR#278)

About this list Date view Thread view Subject view Author view Other groups

Subject: buglet in lm.influence() -- easy fix but... (PR#278)
maechler@stat.math.ethz.ch
Date: Thu 16 Sep 1999 - 01:36:55 EST


Message-Id: <199909151536.RAA18972@pubhealth.ku.dk>

(R 0.65.0 and earlier):

       example(aov)
       lm.influence(npk.aov)
gives
        Error: NA/NaN/Inf in foreign function call (arg 6)

which comes from the lm.obj$coefficients
(the last 3-way interaction coefficient is indetermined and accordingly NA)
argument passed to .Fortran("lminfl",...) inside lm.influence.

The following is a simple correct patch

--- lm.influence.R 1999/08/13 06:44:12 1.5
+++ lm.influence.R 1999/09/15 15:32:11
@@ -38,6 +38,7 @@
              hat = double(n),
              coefficients = matrix(0, nr = n, nc = k),
              sigma = double(n),
+ NAOK = TRUE,# lm.obj$coef can have "structural" NAs
              DUP = FALSE)[c("hat", "coefficients", "sigma")]
 }

However, closer inspection shows that this 6th argument
(called `b' inside src/appl/lminfl.f) is not used at all in the code.

Ross, can I eliminate it from the subroutine lminfl() altogether ?

Martin Maechler <maechler@stat.math.ethz.ch> http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum LEO D10 Leonhardstr. 27
ETH (Federal Inst. Technology) 8092 Zurich SWITZERLAND
phone: x-41-1-632-3408 fax: ...-1228 <><

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._


About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b25 : Tue 04 Jan 2000 - 14:16:08 EST