Seems to me that you could use modulus to simplify the tests.

Your number x (assumed to be less than or equal to 64?)

if x % 8 = 0 it's in the right hand column
if x % 8 = 1 it's in the left hand column

then you just have the ranges

2<=x<=7

and

58<=x<=63

---------- Post updated at 05:31 PM ---------- Previous update was at 05:03 PM ----------

Ahhh...just realized, this test is wrong!

Code:

if ( (res1[1] in set) == (res2[1] in set) );

You can't test for the value of the array element this way, only that the subscript exists!

You could set your array differently instead of using set[i++] why not
break up the array and use set[n]? Then your test should work as you would have elements as follows:

Code:

set[1] through set[8]
set[9], set[16]
set[17], set[24]
set[25], set[32]
set[33], set[40]
set[41], set[48]
set[49], set[56]
set[57] through set[64]

Two other things I noticed...remove the semicolon after the test
and change "==" to &&

Code:

if ( (res1[1] in set) && (res2[1] in set) );

This is the biggest problem your script had other than trying to
use "in" to test the set values instead of the subscripts.

This code worked for me.

Code:

#!/usr/bin/awk -f

 BEGIN  {
   i=0
   for (n=1; n<=8; n++) set[n] = n;
   for (n=9; n<=49; n+=8) {
     set[n] = n 
     set[n+7] = n+7 
   };
   for (n=57; n<=64; n++) set[n] = n;
 }

 ($1 == "|") {
     split($3, res1, "@"); split($6, res2, "@");
     if ( (res1[1] in set) && (res2[1] in set) ) # <--- no ';' here!
     {
       print;
     }

 }

udc1.txt was both your first and second examples put together in that order.