[R] How to read in data

Barry Rowlingson B.Rowlingson at lancaster.ac.uk
Tue Jul 15 17:05:21 CEST 2003


Anne Piotet wrote:
> Hello,
> I'm new to R and in the process of testing it
> My first question: I fail to read in my data (ANSI toto.txt file, tab separated)
>             > test <-read.table("toto.txt")
>             Error in file(file, "r") : unable to open connection
>             In addition: Warning message: 
>             cannot open file `toto.txt'

  - that's because it didnt find the file in that location.

>             > test <- scan("C:\\toto.txt")
>             Error in scan("C:\\toto.txt") : "scan" expected a real, got "No_D"

  - that's because it did find the file, but there was the text "No_D" 
in it. scan() will only read numbers unless you tell it otherwise.

>             > test <-scan("test.dat")
>             Error in file(file, "r") : unable to open connection
>             In addition: Warning message: 
>             cannot open file `toto.txt

  again, its not looking in c:\, so it doesn't find it. Funny how 
scan("test.dat") brings up an error about "toto.txt" :)

  R has a working directory which is where scan() and read.file() will 
start looking for files without a full path - type getwd() to see where 
that is at any time.

  You didnt try the other option:

   test <- read.table("c:\\toto.txt", sep='\t')

- I give a full path to toto.txt and tell it the columns are separated 
with tabs ('\t'). You may need other options - popular ones are as.is=T 
which keeps character variables as text rather than converting to 
categorical data (factors), and head=T if the first line of the file is 
a header with column names.

  If this works, then do names(test) and summary(test) to see what 
you've got.

> second question...what are the size limits of statistical files I can handle? I plan to analize plant datas (up to 500'000 records, from which I will analize a restrictive set of variates ) Even when broken down by some chracteristics, the data to analize can have 50'000-100'000 records

  Depends - whats the size of the machine you are using (and dont say 
its a small box that fits under my monitor). How much RAM and disk space 
does it have?

Baz




More information about the R-help mailing list