Concatenating multiple lines with a discriminator - scripting

Question

I have the input like this
Input:
a,b,c
d,e,f
g,h,i
k,l,m
n,o,p
q,r,s
I wan to be able to concatenate the lines with a discriminator like "|"
Output:
a,b,c|d,e,f|g,h,i
k,l,m|n,o.p|q,r,s
The file has 1million lines and I want to be able to concatenate lines like the example before.
Any ideas about how to approach this?

Answer 1

#OP, if you want to group them for every 3 records,
$ awk 'ORS=(NR%3==0)?"\n":"|"' file
a,b,c|d,e,f|g,h,i
k,l,m|n,o,p|q,r,s
with Perl,
$ perl -lne 'print $_ if $\ = ($. % 3 == 0) ? "\n" : "|"' file
a,b,c|d,e,f|g,h,i
k,l,m|n,o,p|q,r,s

Answer 2

Since your tags include sed here is a way to use it:
sed 'N;N;s/\n/|/g' datafile

Answer 3

gawk:
BEGIN {
state=0
}
state==0 {
line=$0
state=1
next
}
state==1 {
line=line "|" $0
state=2
next
}
state==2 {
print line "|" $0
state=0
next
}

Answer 4

If Perl is fine, you can try:
$i = 1;
while(<>) {
chomp;
unless($i % 3)
{ print "$line\n"; $i = 1; $line = "";}
$line .= "$_|";
$i++;
}
to run:
perl perlfile.pl 1millionlinesfile.txt

Answer 5

$ paste -sd'|' input | sed -re 's/([^|]+\|[^|]+\|[^|]+)\|/\1\n/g'
With paste, we join the lines together, and then sed dices them up. The pattern grabs runs of 3 pipe-terminated fields and replaces their respective final pipes with newlines.
With Perl:
#! /usr/bin/perl -ln
push #a => $_;
if (#a == 3) {
print join "|" => #a;
#a = ();
}
END { print join "|" => #a if #a }