How to override coerion error in 'scan'

Uwe,

Thanks. I did find something in the archives about using a custom
colClass of 'num' with read.csv and using setAs to define a character ->
num function. From the read.table code this appears to force 'scan' to
read it as character and then convert it later using 'as'.  I'm not sure
if there is any advantage in this approach to just reading it using a
colClass of 'character' and then converting it myself afterward (?)

--Rich

Richard Kittler 
AMD TDG
408-749-4099

-----Original Message-----
From: Uwe Ligges [mailto:ligges at statistik.uni-dortmund.de] 
Sent: Saturday, April 23, 2005 6:50 AM
To: Kittler, Richard
Cc: r-help at stat.math.ethz.ch
Subject: Re: [R] How to override coerion error in 'scan'

I am using 'read.csv' in V2.0.1 to read in a CSV file with the 
colClasses option and am getting an error from 'scan' when it 
encounters a non-numeric value for a 'numeric' column, i.e.

 > ds <- read.csv(in_file, nrows=irow, row.names=NULL, 

colClasses=zclass,
                     comment.char="")
  Error in scan(file = file, what = what, sep = sep, quote = quote, 
dec = dec, :
     "scan" expected a real, got "03/15/200523:56:03"

Is there a way to override this and just have it convert those values 
to NA? The dataset is large so I would prefer not to have to import 
the columns as character and convert them to numeric afterward.

I think you have to read it in as character - or write your own C-level 
facility...

Uwe Ligges

--Rich

Richard Kittler
AMD TDG
408-749-4099

______________________________________________
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