[R] Bug in nlme under version 2.0.0

[Douglas Bates]

Søren Højsgaard wrote:
> Dear all,
> Under version 2.0.0, I get the error below when calling summary() on a lme-object, whereas it works under version 1.9.1 (well, it did last week, before I upgraded). Any help on this?
> Thx in advance
> Søren
> 
> 
>>library(nlme)
>>mf <- formula(Weight~Cu*(Time+I(Time^2)+I(Time^3)))
>>lme1 <- lme(mf, data = dietox, random=~1|Pig)
>>summary(lme1)
> 
> Linear mixed-effects model fit by REML
>  Data: dietox 
>        AIC      BIC    logLik
>   4748.664 4815.081 -2360.332
> 
> Random effects:
>  Formula: ~1 | Pig
>         (Intercept) Residual
> StdDev:    6.360083 3.116751
> 
> Fixed effects: Error in as.vector(x, "list") : cannot coerce to vector

I can't reproduce this problem on a simple example like

library(nlme)
example(lme)
summary(fm1)

Can you give us more information on your platform, the version of nlme, 
and perhaps make the data set available on the Internet so we can try to 
reproduce the problem?

Just as a guess, do you happen to have the lme4 package loaded before 
loading nlme?  The two packages cannot be loaded simultaneously.  This 
is detected if you load them in the order nlme, lme4 but not if you load 
them in the other order.