Shell Programming and Scripting

Hi all,
I'm trying to select lines of a file and at the end do not have redundant itens:

Input
A_B
K_A
C_T
A_O
U_B
P_C
D_F
Z_G
W_U

Output
A_B
C_T
D_F
Z_G
W_U

Note: You can see that the output does not have the itens K_A, A_O, U_B and P_C because A, B and C were present at least one time in output. Moreover, the W_U is present because U is combined with W and B, but in the later case the letter B was alread present in the output. In other words, each item must be present just one time in the output.
I tried a lot of things, but I was not able to rigth this script.

Regards

V

Have a go with this:

Code:

awk '
    {
        split( $1, a, "_" );
        if( !printed[a[1]]  && !printed[a[2]] ) 
        {
            printed[a[1]] = printed[a[2]] = 1;
            print;
        }
    }
' input-file >output-file

A much less readable and maintainable version of agama's solution which is guaranteed to make whoever inherits the code mutter under their breath:

Code:

awk -F_ '!a[$1] && !a[$2] && ++a[$1] && ++a[$2]' file

Regards,
Alister