BEGINNER: extracting subsets of data

Question

Hello, this seems to be an easy task really, but being completely new to the world of programming, I have problems with the following task: I have a huge file which has the following format:

track type= wiggle name09
variableStep chrom=chr1
34 5 
36 7 
54 8 
variableStep chrom=chr2 
33 4 
35 2 
78 7 
this is text with the word random in it# this we need to remove
82 4 
88 6 
variableStep chrom=chr3 
78 5 
89 4 
56 7

now what I would like as an out put is just

one file called 1 and containing only

34 5
36 7
54 8

a second file called 2

33 4
35 2
78 7
82 4 
88 6

a third file

78 5
89 4
56 7

It would be great to get some help on this... If any knows how to do it in R... that would be even better

Answer 1

Does the following help?

#!/usr/bin/env perl

use strict;
use warnings;

my $filename = 1;
my $flag;
my $fh;

while (<>) {
    if (/^\d+\s+\d+\s*$/) {
        if ( $flag == 1 ) {
            $flag = 0;
            open $fh, '>', $filename;
            $filename++;
        }
        print $fh $_;
    }
    elsif (/random/) {
        next;
    }
    else {
        $flag = 1;
    }
}

Usage:

Save the above as extract (or any other name, if that matters).

Assuming that the file with data is named file.

perl extract /path/to/file

Answer 2

Here's a solution in R.

Load your data:

a <- readLines(textConnection("track type= wiggle name09
variableStep chrom=chr1
34 5 
36 7 
54 8 
variableStep chrom=chr2 
33 4 
35 2 
78 7 
this is text with the word random in it# this we need to remove
82 4 
88 6 
variableStep chrom=chr3 
78 5 
89 4 
56 7"))

Process it by finding the break points and only keeping rows with number space number format:

idx <- grep("=", a)
idx <- idx[c(which((idx[-1]-idx[-length(idx)])>1),length(idx))]
idx <- cbind(idx+1,c(idx[-1]-1,length(a)))
sapply(1:nrow(idx), function(i) {
 x <- a[idx[i,1]:idx[i,2]]
 write.table(x[grep("^\\d+\\s+\\d+\\s*", x, perl=TRUE)], file=as.character(i), row.names=FALSE, col.names=FALSE, quote=FALSE)
})