Hi all,

I am really in need of some light here. Any suggestions deeply appreciated !
I have a relative simple problem but there is something missing.

This is the task: I have a file with N lines. N is composed of M subgroups. For instance, let's say that each subgroup has 3 lines :

Line 1 : A1 (A1 is actually composed by a few columns)
Line 2 : A2
Line 3 : A3
Line 4 : B1
Line 5 : B2
Line 6 : B3
...
Line N : 3xM

I just need to filter this file and put every n-th line in a different file :

File 1
Line 1 : A1
Line 2 : B1
Line 3 : C1
...

File 2
Line 1 : A2
Line 2 : B2
Line 3 : C2
...

File 3
Line 1 : A3
Line 2 : B3
Line 3 : C3
...

Initially I tried to use "sed -n" like this :

sed -n '1,${p;n;n;}' data > result

This works only for small subgroups like the one above. If data has millions of lines with thousands of subgroups, it is impossible to run sed command (because it has thousands of "n;n;n...").

Does someone know how to perform an efficient selective printing ?
Thanks !

Solved.
For the record, instead of echoing thousands of "n;" in a single line, I included a \n (new line) every hundred or so. Not optimal but it works.

I'm not a sed guy, but I'd do this in perl:

perl -MIO::Handle -e 'open $h[$_], ">", "file$_" for 1..10; $c=1; while(<>) { $h[$c++]->print($_); $c=1 if $c > 10; } close $h[$_] for 1..10' < data

That will split the input over 10 separate files named file1 through file10. Change 10 to whatever number you want.