as.data.frame.matrix() returns an invalid object

identical(as.list(df2),as.list(df1))

Hi,

Two ways to create what should normally be the same data frame:

  > df1 <- data.frame(a=character(0), b=character(0))> df1

  [1] a b
  <0 rows> (or 0-length row.names)

  > df2 <- as.data.frame(matrix(character(0), ncol=2, dimnames=list(NULL,

letters[1:2])))

  > df2

  [1] a b
  <0 rows> (or 0-length row.names)

unique() works as expected except that I get a warning on 'df2':

  > unique(df1)

  [1] a b
  <0 rows> (or 0-length row.names)

  > unique(df2)

  [1] a b
  <0 rows> (or 0-length row.names)
  Warning message:
  In is.na(rows) : is.na() applied to non-(list or vector) of type 'NULL'

Look like the two data frames are not identical:

  > identical(df1, df2)

  [1] FALSE

  > all.equal(df1, df2)

  [1] "Attributes: < Length mismatch: comparison on first 1 components >"

  > attributes(df1)

  $names
  [1] "a" "b"

  $row.names
  integer(0)

  $class
  [1] "data.frame"

  > attributes(df2)

  $names
  [1] "a" "b"

  $class
  [1] "data.frame"

Actually 'df2' is considered broken by validObject():

  > validObject(df1)

  [1] TRUE

  > validObject(df2)

  Error in validObject(df2) :
    invalid class ?data.frame? object: slots in class definition but not in
object: "row.names"

This is with R 2.15 and recent R devel.

Cheers,
H.

--
Herv? Pag?s

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 fhcrc.org
Phone:  (206) 667-5791
Fax:    (206) 667-1319

... and further

identical(as.list(df2),as.list(df1))

[1] TRUE

in R 2.15.0

Not sure whether these sorts of degenerate cases are of much value,
though. But I'll leave that for the wizards.

-- Bert

On Wed, Oct 10, 2012 at 11:22 PM, Herv? Pag?s <hpages at fhcrc.org> wrote:

Hi,

Two ways to create what should normally be the same data frame:

df1 <- data.frame(a=character(0), b=character(0))> df1

 [1] a b
 <0 rows> (or 0-length row.names)

df2 <- as.data.frame(matrix(character(0), ncol=2, dimnames=list(NULL,

letters[1:2])))

df2

 [1] a b
 <0 rows> (or 0-length row.names)

unique() works as expected except that I get a warning on 'df2':

unique(df1)

 [1] a b
 <0 rows> (or 0-length row.names)

unique(df2)

 [1] a b
 <0 rows> (or 0-length row.names)
 Warning message:
 In is.na(rows) : is.na() applied to non-(list or vector) of type 'NULL'

Look like the two data frames are not identical:

identical(df1, df2)

 [1] FALSE

all.equal(df1, df2)

 [1] "Attributes: < Length mismatch: comparison on first 1 components >"

attributes(df1)

 $names
 [1] "a" "b"

 $row.names
 integer(0)

 $class
 [1] "data.frame"

attributes(df2)

 $names
 [1] "a" "b"

 $class
 [1] "data.frame"

Actually 'df2' is considered broken by validObject():

validObject(df1)

 [1] TRUE

validObject(df2)

 Error in validObject(df2) :
   invalid class ?data.frame? object: slots in class definition but not in
object: "row.names"

This is with R 2.15 and recent R devel.

Cheers,
H.

--
Herv? Pag?s

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 fhcrc.org
Phone:  (206) 667-5791
Fax:    (206) 667-1319

-- 

Bert Gunter
Genentech Nonclinical Biostatistics

Internal Contact Info:
Phone: 467-7374
Website:
http://pharmadevelopment.roche.com/index/pdb/pdb-functional-groups/pdb-biostatistics/pdb-ncb-home.htm