R help - Aug 2006 - as.data.frame(cbind()) transforming numeric to factor?

Dear List,

why does as.data.frame(cbind()) transform numeric variables to
factors, once one of the other variablesused is a character vector?

#
x.1 <- rnorm(10)
x.2 <- c(rep("Test",10))
Foo <- as.data.frame(cbind(x.1))
is.factor(Foo$x.1)

Foo <- as.data.frame(cbind(x.1,x.2))
is.factor(Foo$x.1)
#

I assume there is a good reason for this, can somebody explain? Thanks.

Best,
Tom

On Fri, 2006-08-18 at 10:41 -0400, Tom Boonen wrote:
> Dear List,
> 
> why does as.data.frame(cbind()) transform numeric variables to
> factors, once one of the other variablesused is a character vector?
> 
> #
> x.1 <- rnorm(10)
> x.2 <- c(rep("Test",10))
> Foo <- as.data.frame(cbind(x.1))
> is.factor(Foo$x.1)
> 
> Foo <- as.data.frame(cbind(x.1,x.2))
> is.factor(Foo$x.1)
> #
> 
> I assume there is a good reason for this, can somebody explain? Thanks.
> 
> Best,
> Tom
See the Note section of ?cbind, which states:

The method dispatching is not done via UseMethod(), but by C-internal
dispatching. Therefore, there is no need for, e.g., rbind.default.

The dispatch algorithm is described in the source file
(?.../src/main/bind.c?) as

     1. For each argument we get the list of possible class memberships
        from the class attribute.
     2. We inspect each class in turn to see if there is an an
        applicable method.
     3. If we find an applicable method we make sure that it is
        identical to any method determined for prior arguments. If it is
        identical, we proceed, otherwise we immediately drop through to
        the default code.

If you want to combine other objects with data frames, it may be
necessary to coerce them to data frames first. (Note that this algorithm
can result in calling the data frame method if the arguments are all
either data frames or vectors, and this will result in the coercion of
character vectors to factors.)


Thus, note the result of:
> str(cbind(x.1, x.2))
 chr [1:10, 1:2] "-0.265756038510064" "2.13220714034528" ...
 - attr(*, "dimnames")=List of 2
  ..$ : NULL
  ..$ : chr [1:2] "x.1" "x.2"

Since a matrix can only contain a single data type, the numeric vector
is coerced to character.

Then using as.data.frame() coerces the character matrix to factors,
which is the default behavior.

If you want to create a data frame, do it this way:
> str(data.frame(x.1, x.2))
`data.frame':   10 obs. of  2 variables:
 $ x.1: num  -0.266  2.132  2.096 -0.128 -0.466 ...
 $ x.2: Factor w/ 1 level "Test": 1 1 1 1 1 1 1 1 1 1

or if you want to retain the character vector, use I():
> str(data.frame(x.1, I(x.2)))
`data.frame':   10 obs. of  2 variables:
 $ x.1: num  -0.266  2.132  2.096 -0.128 -0.466 ...
 $ x.2:Class 'AsIs'  chr [1:10] "Test" "Test"
"Test" "Test" ...


See ?data.frame for more information.

HTH,

Marc Schwartz

In "R version 2.4.0 Under development (unstable) (2006-08-08 r38825)"
one can do this:

as.data.frame(cbind(x.1,x.2),stringsAsFactors = FALSE)


On 8/18/06, Tom Boonen <tom.boonen.maiden at gmail.com>
wrote:
> Dear List,
>
> why does as.data.frame(cbind()) transform numeric variables to
> factors, once one of the other variablesused is a character vector?
>
> #
> x.1 <- rnorm(10)
> x.2 <- c(rep("Test",10))
> Foo <- as.data.frame(cbind(x.1))
> is.factor(Foo$x.1)
>
> Foo <- as.data.frame(cbind(x.1,x.2))
> is.factor(Foo$x.1)
> #
>
> I assume there is a good reason for this, can somebody explain? Thanks.
>
> Best,
> Tom
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

On Fri, 18 Aug 2006, Tom Boonen wrote:
> Dear List,
> 
> why does as.data.frame(cbind()) transform numeric variables to
> factors, once one of the other variablesused is a character vector?
> 
> #
> x.1 <- rnorm(10)
> x.2 <- c(rep("Test",10))
> Foo <- as.data.frame(cbind(x.1))
> is.factor(Foo$x.1)
> 
> Foo <- as.data.frame(cbind(x.1,x.2))
> is.factor(Foo$x.1)
> #
> 
> I assume there is a good reason for this, can somebody explain? Thanks.
Only if you can explain the good reason why you did not just use 
data.frame(x.1, x.2)!

cbind() makes a matrix out of vectors, here a character matrix.  And then 
as.data.frame() converts character columns to factors.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

Thanks everybody. I recognize my mistake now.

I think as.data.frame(cbind(x.1,x.2),stringsAsFactors = FALSE) would
be a good idea.

Tom

On 8/18/06, Prof Brian Ripley <ripley at stats.ox.ac.uk>
wrote:
> On Fri, 18 Aug 2006, Tom Boonen wrote:
>
> > Dear List,
> >
> > why does as.data.frame(cbind()) transform numeric variables to
> > factors, once one of the other variablesused is a character vector?
> >
> > #
> > x.1 <- rnorm(10)
> > x.2 <- c(rep("Test",10))
> > Foo <- as.data.frame(cbind(x.1))
> > is.factor(Foo$x.1)
> >
> > Foo <- as.data.frame(cbind(x.1,x.2))
> > is.factor(Foo$x.1)
> > #
> >
> > I assume there is a good reason for this, can somebody explain?
Thanks.
>
> Only if you can explain the good reason why you did not just use
> data.frame(x.1, x.2)!
>
> cbind() makes a matrix out of vectors, here a character matrix.  And then
> as.data.frame() converts character columns to factors.
>
> --
> Brian D. Ripley,                  ripley at stats.ox.ac.uk
> Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
> University of Oxford,             Tel:  +44 1865 272861 (self)
> 1 South Parks Road,                     +44 1865 272866 (PA)
> Oxford OX1 3TG, UK                Fax:  +44 1865 272595
>

>>>>> "Tom" == Tom Boonen <tom.boonen.maiden at
gmail.com>
>>>>>     on Fri, 18 Aug 2006 11:16:45 -0400 writes:
    Tom> Thanks everybody. I recognize my mistake now.
    Tom> I think    as.data.frame(cbind(x.1,x.2),stringsAsFactors = FALSE)
    Tom> would be a good idea.

I think

	data.frame(x.1, x.2 = I(x.2))

would be a considerably better idea.

[ The use of I(.) for preventing coercion to factors 
  is a much older and "S-like" way ]

Martin



    Tom> Tom

    Tom> On 8/18/06, Prof Brian Ripley <ripley at stats.ox.ac.uk>
wrote:
    >> On Fri, 18 Aug 2006, Tom Boonen wrote:
    >> 
    >> > Dear List,
    >> >
    >> > why does as.data.frame(cbind()) transform numeric variables to
    >> > factors, once one of the other variablesused is a character
vector?
    >> >
    >> > #
    >> > x.1 <- rnorm(10)
    >> > x.2 <- c(rep("Test",10))
    >> > Foo <- as.data.frame(cbind(x.1))
    >> > is.factor(Foo$x.1)
    >> >
    >> > Foo <- as.data.frame(cbind(x.1,x.2))
    >> > is.factor(Foo$x.1)
    >> > #
    >> >
    >> > I assume there is a good reason for this, can somebody
explain? Thanks.
    >> 
    >> Only if you can explain the good reason why you did not just use
    >> data.frame(x.1, x.2)!
    >> 
    >> cbind() makes a matrix out of vectors, here a character matrix. 
And then
    >> as.data.frame() converts character columns to factors.
    >> 
    >> --
    >> Brian D. Ripley,                  ripley at stats.ox.ac.uk
    >> Professor of Applied Statistics, 
http://www.stats.ox.ac.uk/~ripley/
    >> University of Oxford,             Tel:  +44 1865 272861 (self)
    >> 1 South Parks Road,                     +44 1865 272866 (PA)
    >> Oxford OX1 3TG, UK                Fax:  +44 1865 272595
    >> 

    Tom> ______________________________________________
    Tom> R-help at stat.math.ethz.ch mailing list
    Tom> https://stat.ethz.ch/mailman/listinfo/r-help
    Tom> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
    Tom> and provide commented, minimal, self-contained, reproducible code.