Laurent Gautier wrote: > Dear list, > > > When calling R from C, what appears like a spurious error can be > triggered during the execution of chisq.test(x, y). > > This is happening when the following conditions are met: > > - x and y are "anonymous" C-level R vectors (they do not have a symbol), > but they are protected from garbage collection > > - x and y are "not too small" (it was experienced as soon as they are > longer than 17 elements). > > > The error is > > Error in names(dimnames(x)) <- DNAME : > 'names' attribute [4] must be the same length as the vector [2] > > and can be traced to the use of deparse(substitute(x)) and > deparse(substitute(y)) in the R code. > > What seems to be happening is that the deparse(substitute(x)) call > gives a character vector of length > 1 as soon as x is "not so small". > > To demonstrate this, I am using rpy2 (as the problem was found when > using it > http://sourceforge.net/mailarchive/forum.php?thread_name=4B043CA1.9050500%40salilab.org&forum_name=rpy-list > > ), but it will likely be present in other bridges to R as well. > > > #using R-2.10 and rpy2-2.1.dev >>>> import rpy2.robjects as robjects > >>>> f = robjects.r('''function(x) return(deparse(substitute(x)))''') > >>>> tuple(f(robjects.FloatVector(range(17)))) > ('c(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)',) > # length 1 >>>> tuple(f(robjects.FloatVector(range(18)))) > ('c(0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17', ')') > # length 2 !! > > > Does it seem to others like an issue present in other bridges as well ? >
It can be generated from R itself: > x <- sample(0:1,2000,replace=T) > y <- sample(0:1,2000,replace=T) > eval(substitute(chisq.test(x,y),list(x=x,y=y))) Error in names(dimnames(x)) <- DNAME : 'names' attribute [252] must be the same length as the vector [2] Looks like nlines=1 is needed in the deparse(). -- O__ ---- Peter Dalgaard Øster Farimagsgade 5, Entr.B c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~~~~~~~~~ - (p.dalga...@biostat.ku.dk) FAX: (+45) 35327907 ______________________________________________ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
