repeat data at one line and grep as pre-set format

[benlaw]

Hi all, I have data like this

00000001  99888  00000011  99889  00000002  99890  00000003  99808  (m
+any many...)   
00000021  99888  00000021  99889  00000022  99890  00000023  99900
(many many)

I would like grep like

00000001  99888
00000011  99889
00000002  99890  
....

How can i do that, Thanks a lot!

[davido]

I'm not entirely sure what you're asking for, but it looks like you're just needing to reformat the data to a two-column output. If that's the case, this ought to do the trick:

while( my $line = <DATA> ) {
    while( $line =~ m/(\d{8})\s+(\d{5})/g ) {
        print "$1 $2\n";
    }
}

If I've missed the point to the question, you might help out by providing further explanation on what you are trying to do.

[demerphq]

Imo if you are going to do that then you should just do an s///

perl -pe "s/(\d{8}\s+\d{5})(\s+)/$1\n/g" t.txt

In sure there is a way to do that without needing to copy $1, but I couldnt work it out quickly enough to post.

[davido]

You could avoid copying $1 if you use lookbehind, but there is that pesky problem of \s+ between the digit fields, which won't work with lookbehind (variable width rule). I couldn't tell by the OP's post whether there was a fixed number of spaces or not. Assuming only two spaces (or at least a known quantity of spaces) between the column fields, you could do something like this:

perl -pe "s/(?<=\d{8}\s\s\d{5})\s+/\n/g;" t.txt

[demerphq]

perl -aF/\s+/ -lne"print join qq(\t),splice(@F,0,2) while @F" t.txt

Adjust the quotes to suite your shell.

[benlaw]

Great thanks!!

[demerphq]

Its not at all clear what you meant by "grep" in this context. Nothing is being grepped.

[saintmike]

Are you saying you want the first two fields extracted from every line? If so, just use split:

while(<DATA>) {
    my @two = split ' ', $_;
    print "@two[0,1]\n";
}
__DATA__
00000001  99888  00000011  99889  00000002  99890  00000003  99808
00000021  99888  00000021  99889  00000022  99890  00000023  99900

[svenXY]

Hi,


perl -n -e'print join(" ", (split(/\s+/))[0,1]), "\n"' yourfile.txt


Regards,
svenXY
Update: fixed the line, works now

[Random_Walk]

I think you want to split each pair of data points out to a line on its own instead of having multiple datapoints on one line.

cat yourfile.txt | perl -ple 's/(\S+\s+\S+)\s+/$1\n/g'

should do this for you as long as you have even numbers of datapoints on each line

echo "00000001  99888  00000011  99889  00000002  99890  00000003  998
+08" \
| perl -ple 's/(\S+\s+\S+)\s+/$1\n/g'
00000001  99888
00000011  99889
00000002  99890
00000003  99808

Cheers,
R.

[davido]

cat yourfile.txt | perl -ple 's/(\S+\s+\S+)\s+/$1\n/g'

Isn't that the same as:

perl -ple 's/(\S+\s+\S+)\s+/$1\n/g' yourfile.txt

UUOC.

[Random_Walk]

You are right and your way is more efficient too. I wrote it with the cat file | as that is how I tested it with echo, just failed to engage my brain completely :-)

Cheers,
R.