This happens in as.matrix(), which gets called by apply().
When you've got a mixed-mode dataframe like this, as.matrix() converts
everything to character. But, the rules it uses for each column don't
seem to be entirely consistent regarding whether columns are
space-padded to make each element have the same number of characters.
The way it works out, logical mode columns are passed through format(),
which space-pads by default. Factor columns are passed through
as.vector(), and character-mode columns are left alone. The result is
that some columns come out space-padded, and some don't, depending on
their original mode.
To get greater control of this, you theoretically should be able to do
something like apply(as.matrix(format(X, justify="none")), ...),
except that format() seems to ignore the justify argument for logical
vectors, e.g.:
> format(c(T,F,T))
[1] " TRUE" "FALSE" " TRUE"
> format(c(T,F,T), justify="none")
[1] " TRUE" "FALSE" " TRUE"
>
If it's really important for you to get this to work the way you want,
you can convert the logical column of the data frame using as.character
(see the end of the example below).
Here's an example that shows probably far more than you wanted to know:
> X <- data.frame(letters=letters[1:3], flag=c(TRUE, FALSE, TRUE),
codef=c("a","ab","abcd"), codec=I(c("x", "xy", "xyz")))
> sapply(X, class)
letters flag codef codec
"factor" "logical" "factor" "AsIs"
> as.matrix(X)
letters flag codef codec
[1,] "a" " TRUE" "a" "x"
[2,] "b" "FALSE" "ab" "xy"
[3,] "c" " TRUE" "abcd" "xyz"
> unclass(format(X))
$letters
[1] "a" "b" "c"
$flag
[1] " TRUE" "FALSE" " TRUE"
$codef
[1] "a" "ab" "abcd"
$codec
[1] "x" "xy" "xyz"
attr(,"row.names")
[1] "1" "2" "3"
> unclass(format(X, justify="left"))
$letters
[1] "a" "b" "c"
$flag
[1] " TRUE" "FALSE" " TRUE"
$codef
[1] "a " "ab " "abcd"
$codec
[1] "x " "xy " "xyz"
attr(,"row.names")
[1] "1" "2" "3"
>
> # The only way I can see to get the logical column converted to
character without padding:
> X1 <- X
> X1$flag <- as.character(X1$flag)
> as.matrix(X1)
letters flag codef codec
[1,] "a" "TRUE" "a" "x"
[2,] "b" "FALSE" "ab" "xy"
[3,] "c" "TRUE" "abcd" "xyz"
>
Adrian Dragulescu wrote:
Hello,
X <- data.frame(letters=letters[1:3], flag=c(TRUE, FALSE, TRUE))
X
letters flag
1 a TRUE
2 b FALSE
3 c TRUE
apply(X, 1, as.list)
[[1]]
[[1]]$letters
[1] "a"
[[1]]$flag
[1] " TRUE"
[[2]]
[[2]]$letters
[1] "b"
[[2]]$flag
[1] "FALSE"
[[3]]
[[3]]$letters
[1] "c"
[[3]]$flag
[1] " TRUE"
Notice how TRUE becomes " TRUE" and FALSE becomes "FALSE". Not sure
why TRUE gets an extra whitespace in front.
Checked with R-2.10.0, but can reproduce the behavior as far back as
R-2.8.1.
Adrian Dragulescu
sessionInfo()
R version 2.10.0 (2009-10-26)
i386-pc-mingw32
locale:
[1] LC_COLLATE=English_United States.1252 LC_CTYPE=English_United
States.1252
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C
[5] LC_TIME=English_United States.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
loaded via a namespace (and not attached):
[1] tools_2.10.0
______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
