[R] RE: packaged datasets in .csv format (David Firth)

Uwe Ligges ligges at statistik.uni-dortmund.de
Thu Jul 10 13:25:32 CEST 2003

Andreas Christmann wrote:
>> ----------------------------------------------------------------------
>> Message: 1
>> Date: Wed, 9 Jul 2003 10:53:27 +0100
>> From: David Firth <david.firth at nuffield.oxford.ac.uk>
>> Subject: [R] packaged datasets in .csv format
>> To: r-help at stat.math.ethz.ch
>> Message-ID:
>>     <307D34CE-B1F3-11D7-A8D2-0050E4C03977 at nuffield.oxford.ac.uk>
>> Content-Type: text/plain; charset=US-ASCII; format=flowed
>> A couple of questions in connection with using .csv format to include 
>> data in a package:
>> First, the background.  The data() function loads data from .csv 
>> ("comma-separated values") files using
>>    read.table(..., header = TRUE, sep = ";")
>> But ?read.table says
>>       ## To write a CSV file for input to Excel one might use
>>       write.table(x, file = "foo.csv", sep = ",", col.names = NA)
>>       ## and to read this file back into R one needs
>>       read.table("file.csv", header = TRUE, sep = ",", row.names=1)
>> As a result, .csv files created by write.table() as above are not read 
>> in by data() in the way that might be expected [that is, expected by 
>> someone who had not read help(data)!]
>> Two questions, then:
>> -- is there some compelling reason for  the use of `sep = ";"' in 
>> place of `sep = ",", row.names=1'?

Do you really want an answer?
Today, one reason is compatibility to all the other packages on CRAN.

> I prefer ";" instead of "," , because in text variables there are often 
> ",".

That's why text variables can be quoted.

>> -- if I want to maintain a dataset in .csv format, for use both in R 
>> and in other systems such as Excel, SPSS, etc, what is the best way to 
>> go about it?

When regularly using that many systems on the same data sets, it might 
be worth using a database system, e.g. MySQL.

BTW: R *and* Excel *and* (for sure, but I haven't tested) also SPSS can 
read a couple of different ASCII formatted files, so there are quite a 
lot possible formats.

Uwe Ligges

> Depends. Perhaps it is best to check it out for the software packages
> and the versions of the software packages you are using.
> Andreas Christmann
>> Any advice would be much appreciated.
>> Cheers,
>> David

More information about the R-help mailing list