Sorry for the mistakes. Now, I repaired the posting.

Hi power user,

Basically, this thread is a continuation of the previous one :

Making Connection nodes for Graph

However, I'm going to explain it again.

I have this following data:

file1
aa A
aa B
aa C
bb X
bb Y
bb Z
cc O
cc P
cc Q
. .
. .
. .
. .

and I want to turn them into a connection nodes like this:
file2

A aa A
A aa B
A aa C
B aa C
B aa B
C aa C
X bb X
X bb Y
X bb Z
Y bb Z
Y bb Y
Z bb Z
. . .
. . .
. . .
. . .

I made this relation, to create a graph. The file have more than 6.000.000 lines.
For smaller files (100.000 lines), I have used this following script in the previous thread:

join -o 1.2 0 2.2 -1 1 -2 1 file1 file1 | nawk '!a[$3$2$1];{a[$1$2$3]++}'

join -o 1.2 0 2.2 -1 1 -2 1 file1 file1 | nawk '$1<$3{print;next}{print$3,$2,$1}' | sort -u

nawk '
NR==FNR { c = a[$1]; a[$1] = c?c" "$2:$2; next }
{ c = a[$1]
if (c) {
split(c,b)
for (k in b) {
p = $2<b[k]?$2" "$1" "b[k]:b[k]" "$1" "$2
if (!d[p]++) print p
}
}
}
' file1 file1