>>>>> "GB" == Göran Broström <[EMAIL PROTECTED]> >>>>> on Wed, 28 Apr 2004 16:00:17 +0200 writes:
<.........> GB> so, >> model.matrix(x ~ y) GB> (Intercept) 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 1 10 1 GB> attr(,"assign") [1] 0 GB> but >> model.matrix(x ~ as.vector(y)) GB> (Intercept) as.vector(y) 1 1 -0.853357506 2 1 GB> -0.711872147 3 1 -0.228785137 4 1 -0.449739758 5 1 GB> 0.173914266 6 1 -0.138766243 7 1 -0.433799800 8 1 GB> 0.234183701 9 1 0.002728104 10 1 0.733590165 GB> attr(,"assign") [1] 0 1 GB> AND >> rr <- model.matrix.default(x ~ y) GB> Segmenteringsfel [EMAIL PROTECTED]:~$ <..............> GB> (Don't ask me what's going on: but 'Segmenteringsfel' GB> means 'Segmentation fault':) Thank you, Göran. Yes, the bug is in model.matrix.default(); one part of the problems happens in the line ans <- .Internal(model.matrix(t, data)) (which only returns the intercept part). How to fix this is not yet clear to me, since we have to decide if the internal C code should do more checking or the R code in model.matrix.default. BTW, a very simple reproducible example is x <- 1:7; y. <- x ; y <- array(y, 7) model.matrix(x ~ y) # which behaves badly when called repeatedly; # which for me means memory allocation problems ## as opposed to model.matrix(x ~ y.) # which is all fine ---- I'll also file this as bug report. Martin ______________________________________________ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html