R help - May 2006 - retaining character matrices when combining into data frames

Hello everyone,

If I want to convert or combine a (large) character matrix into a data
frame without having any of its columns convert into a factor class,
is there a simple solution? I() says it will operate on 'an
object' but it seems that unless the object is a vector, the results
are not what I expect. 

For instance, if g is a 2x2 character matrix, as.data.frame(I(g)) will
return an object of the data frame class but not structured in the way
I intended. What I hope to retrieve is the result of

h = data.frame(I(g[,1]),I(g[,2]))
names(h) = dimnames(g)[[2]]

With data sets of 100+ columns, wrapping I() around each column can be
very time-consuming when invoking the functions data.frame(), cbind(),
or as.data.frame(). R used to offer the optional argument, as.is=TRUE
(and S-PLUS offered stringsAsFactors=FALSE), to accomplish what this
task but it seems that this argument was removed some time ago. Is
there a more attractive alternative available now?

Thanks very much,

Satoshi


__________
Satoshi Takahama
Scripps Institution of Oceanography
Center for Atmospheric Science
9500 Gilman Drive, Dept. 0221
La Jolla, CA 92093
858-531-5328

Try this:

# test data
mat <- matrix(letters, 2)

# convert to data frame with character columns
DF <- replace(as.data.frame(mat),,mat)

On 5/15/06, Satoshi Takahama <stakahama at ucsd.edu>
wrote:
> Hello everyone,
>
> If I want to convert or combine a (large) character matrix into a data
> frame without having any of its columns convert into a factor class,
> is there a simple solution? I() says it will operate on 'an
> object' but it seems that unless the object is a vector, the results
> are not what I expect.
>
> For instance, if g is a 2x2 character matrix, as.data.frame(I(g)) will
> return an object of the data frame class but not structured in the way
> I intended. What I hope to retrieve is the result of
>
> h = data.frame(I(g[,1]),I(g[,2]))
> names(h) = dimnames(g)[[2]]
>
> With data sets of 100+ columns, wrapping I() around each column can be
> very time-consuming when invoking the functions data.frame(), cbind(),
> or as.data.frame(). R used to offer the optional argument, as.is=TRUE
> (and S-PLUS offered stringsAsFactors=FALSE), to accomplish what this
> task but it seems that this argument was removed some time ago. Is
> there a more attractive alternative available now?
>
> Thanks very much,
>
> Satoshi
>
>
> __________
> Satoshi Takahama
> Scripps Institution of Oceanography
> Center for Atmospheric Science
> 9500 Gilman Drive, Dept. 0221
> La Jolla, CA 92093
> 858-531-5328
>
> ______________________________________________
> 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
>

On Mon, 15 May 2006, Satoshi Takahama wrote:
> Hello everyone,
>
> If I want to convert or combine a (large) character matrix into a data
> frame without having any of its columns convert into a factor class,
> is there a simple solution? I() says it will operate on 'an
> object' but it seems that unless the object is a vector, the results
> are not what I expect.
>
> For instance, if g is a 2x2 character matrix, as.data.frame(I(g)) will
> return an object of the data frame class but not structured in the way
> I intended.
Which is?  Your subject line says you want to retain a character matrix, 
and that is exactly what happens:
> g <- matrix(letters[1:4], 2, 2)
> as.data.frame(I(g))
   x.1 x.2
1   a   c
2   b   d
> What I hope to retrieve is the result of
>
> h = data.frame(I(g[,1]),I(g[,2]))
> names(h) = dimnames(g)[[2]]
So it seems you don't actually want to `retain character matrices'.
> With data sets of 100+ columns, wrapping I() around each column can be
> very time-consuming when invoking the functions data.frame(), cbind(),
> or as.data.frame().
Really?  How are you trying to do it?
> R used to offer the optional argument, as.is=TRUE
> (and S-PLUS offered stringsAsFactors=FALSE), to accomplish what this
> task but it seems that this argument was removed some time ago.
It was not there in R 1.0.0, so you are talking about alpha/beta versions.
I think it existed only when data.frame was a rather different class.
> Is there a more attractive alternative available now?
I guess what you want is for each column of a character matrix to be 
inserted as a character vector.  There are several ways: a simple one is

DF <- data.frame(g)
for(i in 1:ncol(g)) DF[[i]] <- g[, i]

-- 
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