R help - Jan 2010 - read.spss: option "to.data.frame" and string variables

Dear R-users,

I am using R version 2.10.1 and package foreign version 0.8-39 under windows.

When reading .sav-Files (PASW Statistics 18.0.1) containing string variables,
these are automatically converted to factors when using option
"to.data.frame = TRUE" (see example below).
It's clear to me why this happens (the default behaviour of a call to
as.data.frame). But this is not always what one might want (or even be aware
of).

So maybe one of the following improvements could be made?
* Add a description of this behaviour in ?read.spss.
* Or (even better): Add an extra argument, like:
read.spss("C:\\temp\\test.sav", to.data.frame = TRUE, stringsAsFactors
= FALSE).

Just a suggestion;
kind regards
Heinrich.

# EXAMPLE:
Suppose there is a simple file "test.sav", containing one variable
("x") of type STRING with 3 values (a,b,c).
> library(foreign)
> test <- read.spss("C:\\temp\\test.sav")
> test
$x
[1] "a       " "b       " "c       "

attr(,"label.table")
attr(,"label.table")$x
NULL

attr(,"codepage")
[1] 1252
> is.factor(test$x)
[1] FALSE
> is.character(test$x)
[1] TRUE
# Ok, that's just fine. But things change when using option
"to.data.frame = TRUE":
> test <- read.spss("C:\\temp\\test.sav", to.data.frame = TRUE)
> test
         x
1 a
2 b
3 c
> is.factor(test$x)
[1] TRUE
> is.character(test$x)
[1] FALSE

It would be an significant undertaking to annotate all the places  
where the default behavior of strings-to-factors conversion might trip  
up the unwary. You are not the first by any means to complain. You  
might:
a) take the step that the Mayo Clinic has taken of setting the default  
in options() to FALSE, or
b) make your own read.spss with your desired arguments, and then put  
it in your .Rprofile.

-- 
David

On Jan 12, 2010, at 5:28 AM, RINNER Heinrich wrote:
> Dear R-users,
>
> I am using R version 2.10.1 and package foreign version 0.8-39 under  
> windows.
>
> When reading .sav-Files (PASW Statistics 18.0.1) containing string  
> variables, these are automatically converted to factors when using  
> option "to.data.frame = TRUE" (see example below).
> It's clear to me why this happens (the default behaviour of a call  
> to as.data.frame). But this is not always what one might want (or  
> even be aware of).
>
> So maybe one of the following improvements could be made?
> * Add a description of this behaviour in ?read.spss.
> * Or (even better): Add an extra argument, like: read.spss("C:\\temp\ 
> \test.sav", to.data.frame = TRUE, stringsAsFactors = FALSE).
>
> Just a suggestion;
> kind regards
> Heinrich.
>
> # EXAMPLE:
> Suppose there is a simple file "test.sav", containing one
variable
> ("x") of type STRING with 3 values (a,b,c).
>> library(foreign)
>> test <- read.spss("C:\\temp\\test.sav")
>> test
> $x
> [1] "a       " "b       " "c       "
>
> attr(,"label.table")
> attr(,"label.table")$x
> NULL
>
> attr(,"codepage")
> [1] 1252
>> is.factor(test$x)
> [1] FALSE
>> is.character(test$x)
> [1] TRUE
> # Ok, that's just fine. But things change when using option  
> "to.data.frame = TRUE":
>> test <- read.spss("C:\\temp\\test.sav", to.data.frame =
TRUE)
>> test
>         x
> 1 a
> 2 b
> 3 c
>> is.factor(test$x)
> [1] TRUE
>> is.character(test$x)
> [1] FALSE
>
> ______________________________________________
> R-help at r-project.org 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.