Linux Parsing Output and Combine Lines - regex

Question

I have the following output (in linux):
wwn = 5001248018b6d7af
node_wwn = 5001248018b6d7ae
wwn = 5001248118b6d7af
node_wwn = 5001248118b6d7ae
I need to parse the output so it will looks like that:
50:01:24:80:18:b6:d7:ae:50:01:24:80:18:b6:d7:af
50:01:24:81:18:b6:d7:ae:50:01:24:81:18:b6:d7:af
the output is the node_wwn and its wwn.
I done this job using some C# app but its too complicated to take the output and past it to the program.
string command = "cat /proc/sll/info |grep -B 9 up| grep wwn";
string temp;
string []commandOutput;
string[] wwns;
string[] wwpns;
hbaList = new List<HBA>();
commandOutput = (exec.RunCommand(command)).Trim().Split('\n');
wwns = Array.FindAll(commandOutput, item => item.StartsWith("wwn = "));
wwpns = Array.FindAll(commandOutput, item => item.StartsWith("node_wwn = "));
for (int i =0; i<wwns.Length;i++)
{
hba = new HBA();
temp = string.Format("{1}", (wwpns[i].Split('='))).Trim();
temp += string.Format("{1}", (wwns[i].Split('='))).Trim();
temp = Regex.Replace(temp, ".{2}", "$0:");
temp = temp.Remove(temp.Length - 1);
hba.ID = temp;
hbaList.Add(hba);
}
Please help...
Thanks Almog

Answer 1

Code writing service, hello :)
I did it with GNU awk.
#!/usr/bin/env awk -f
function out_wwn(wwn, node_wwn) {
out = sprintf("%s%s", node_wwn, wwn);
patsplit(out, a, "[0-9a-f]{2}");
for (i = 1; i <= length(a); i++) {
if (i > 1) printf(":");
printf("%s", a[i]);
}
printf("\n");
}
/^wwn/ {wwn=$3}
/^node_wwn/ {out_wwn(wwn, $3)}
Usage in your case:
cat /proc/sll/info |grep -B 9 up| awk -f [the file above]
Maybe you can also remove the first grep, depending on your input. In that case you need just do:
awk -f [the file above] < /proc/sll/info

Answer 2

Guessing a bit here...
awk '$1 == "wwn" { result = $3; next }
$1 == "node_wwn" { result = $3 result; next }
/up/ { for(i=1, sep = ""; i<length(result); sep = ":", i += 2)
printf("%s%s", sep, substr(result, i, 2));
printf "\n" }' /proc/sll/info
Assuming the wwn will always precede the node_wwn and be followed by up before the next output record, and that whitespace patterns are consistent.

Answer 3

Here's a perl version
while (<>) {
if ($_ =~ /(node_)*wwn\s*=\s*(\S+)/) {
$node = "$2$node";
if($1 eq "node_") {
$node =~ s/(..)(?!$)/\1:/g;
print "$node\n";
$node = "";
}
}
}

Answer 4

This might work for you (GNU sed):
sed -r 'N;s/.* = (\S+).* = (\S+).*/\2\1/;s/..\B/&:/g' file
This reads two lines at time, removes the cruft and re-arranges the data then inserts a : between each two characters followed by a non-word boundary.

Answer 5

Just throwing out another option.
This assumes node_wwn line directly follows the wwn line..
$ cat file
wwn = 5001248018b6d7af
node_wwn = 5001248018b6d7ae
wwn = 5001248118b6d7af
node_wwn = 5001248118b6d7ae
$ awk '$1=="wwn"{nf=$NF;getline;print $NF""nf}' file | perl -pe 's/\w{2}\B/$&:/g'
50:01:24:80:18:b6:d7:ae:50:01:24:80:18:b6:d7:af
50:01:24:81:18:b6:d7:ae:50:01:24:81:18:b6:d7:af