mjhjmtu

I have a big file with many columns and rows, looks like:

A B C D E F1 F2 F3 F4 F5
a1 b1 c1 d1 e1 0 0 1 0 1
a2 b2 c2 d2 e2 1 0 0 1 1
a3 b3 c3 d3 e3 1 1 0 0 1
....

The A, B, C, D, E columns contain some information, and F1-5 columns represent some ids. The 0s or 1s mean absence/presence of the A-E information for that id.

I want to create files for each id, while every file contains the ABCDE information that the id has.
For example, F5 have three 1s in the first 3 rows, so

F5.txt:

A B C D E 
a1 b1 c1 d1 e1 
a2 b2 c2 d2 e2 
a3 b3 c3 d3 e3

F1 has two 1s in the first 3 rows, so

F1.txt:

A B C D E 
a2 b2 c2 d2 e2 
a3 b3 c3 d3 e3

How to filter this file and create new files with the id names (F1, F2...) using awk?

AWK solution:

awk 'NR==1 split($0,h); columns=sprintf("%s %s %s %s %s",h[1],h[2],h[3],h[4],h[5]); next 
 for (i=6;i<=NF;i++) 
 if ($i) 
 if (!a[h[i]]++) print columns > h[i]".txt"; 
 print $1,$2,$3,$4,$5 > h[i]".txt" 
 
 ' file

• split($0,h) - split the 1st record into array h to obtain header column names




• columns=sprintf("%s %s %s %s %s",h[1],h[2],h[3],h[4],h[5]) - constructing common columns string A B C D E




• if($i) - if the current field (starting from the 6th field) is not empty, i.e. not ""(empty string) or 0 - ready for further processing




• h[i] - points to the current filename, i.e. F1 etc (or as you wrote: represent some ids)




• if (!a[h[i]]++) print columns > h[i]".txt" - if the file under name h[i] is first time written - print the header/columns line to it (as 1st line)

Viewing results:

$ head F*.txt
==> F1.txt <==
A B C D E
a2 b2 c2 d2 e2
a3 b3 c3 d3 e3

==> F2.txt <==
A B C D E
a3 b3 c3 d3 e3

==> F3.txt <==
A B C D E
a1 b1 c1 d1 e1

==> F4.txt <==
A B C D E
a2 b2 c2 d2 e2

==> F5.txt <==
A B C D E
a1 b1 c1 d1 e1
a2 b2 c2 d2 e2
a3 b3 c3 d3 e3