R help - Sep 2006 - read csv

Hi,

I have a csv file where the number of filled columns varies in the
different rows:

Sun 5-Feb-06,15,,,01:30:00,0:06:00,
Mon 6-Feb-06,,,,,,
Tue 7-Feb-06,7,,,00:41:00,0:05:51,
Wed 8-Feb-06,,,,,,

I would like to use read.table (or whatever is appropriate) to read in
only those rows that have two or more columns filled. Any hint will be
appreciated.

Georg

On Mon, 4 Sep 2006, Georg Otto wrote:
> 
> Hi,
> 
> I have a csv file where the number of filled columns varies in the
> different rows:
> 
> Sun 5-Feb-06,15,,,01:30:00,0:06:00,
> Mon 6-Feb-06,,,,,,
> Tue 7-Feb-06,7,,,00:41:00,0:05:51,
> Wed 8-Feb-06,,,,,,
> 
> I would like to use read.table (or whatever is appropriate) to read in
> only those rows that have two or more columns filled. Any hint will be
> appreciated.
See ?read.table is the main hint.

Use read.table(fill=TRUE) and post-process, e.g. by

A <- A[rowSums(!is.na(A)) > 2), ]

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

> I have a csv file where the number of filled columns
> varies in the different rows:
I would hack it like this, but then I am totally new to R, 
which means you should not trust me.:

d <- read.csv("testdata", header=F)

selection <- apply(d, c(1), 
      function(x) {sum(!is.na(x) & x != "") > 2})

as.data.frame(t(as.data.frame(t(d))[selection]))


Tom