[Rd] collapsing list of NULLs dimnames into a NULL

[Hervé Pagès]

Hi,

Here are two different ways to create a 0 x 0 logical matrix:

   m1 <- matrix(nrow=0, ncol=0)
   m1
   # <0 x 0 matrix>

   m2 <- as.matrix(data.frame())
   m2
   # <0 x 0 matrix>

Surprisingly:

   identical(m1, m2)
   # [1] FALSE

That's because of their dimnames:

   dimnames(m1)
   # NULL

   dimnames(m2)
   # [[1]]
   # NULL
   #
   # [[2]]
   # NULL

More generally this can be observed with 0-col matrices:

   m1 <- matrix(nrow=5, ncol=0)
   m2 <- as.matrix(data.frame(stuff=11:15)[0])
   identical(m1, m2)  # FALSE

The culprit is this line in as.matrix.data.frame():

   dimnames(X) <- list(dn[[1L]], unlist(collabs, use.names = FALSE))

Shouldn't as.matrix.data.frame() collapse the list of NULLs into a NULL
before setting the dimnames?

Alternatively, maybe a more general solution could be to have the
dimnames() setter take care of this. There seems to be a lot of
code around that is potentially setting a list of NULLs as the dimnames
of an object, so having dimnames<- do the collapse would fix the problem
in one place once for all and avoid a lot of repeated code.

Finally, having this functionality available at the C level would be
great. Would basically be a wrapper to

   Rf_setAttrib(x, R_DimNamesSymbol, dimnames)

that takes care of the collapse. The SET_DIMNAMES macro could be
redefined to call that wrapper instead...

Thanks,
H.

-- 
Hervé Pagès

Program in Computational Biology
Division of Public Health Sciences
Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N, M1-B514
P.O. Box 19024
Seattle, WA 98109-1024

E-mail: hpages at fredhutch.org
Phone:  (206) 667-5791
Fax:    (206) 667-1319