How to write the regular expression to extract these fields from my sample data?

[kschmeling]

Hello,

I'm trying to create a regex to extract the fields to the follow logs:

Example 1

msg=O equipamento marte (192.168.0.1) recuperou a conectividade SSH suid=7 sname=Script Assíncrono: 10 suser= spid=17025 dst=192.168.0.2 dhost=marte dport=22

Example 2

msg=Sessao iniciada para marte (192.168.0.1) - Usuario Privilegiado TPL - root pelo usuario Usuario Teste (usuario.teste) suid=297 sname=Usuario Teste suser=usuario.teste spid=14410 dst=192.168.0.1 dpt=22 duser=root

I need to extract the fields like that:

Example 1

msg=O equipamento marte (192.168.0.1) recuperou a conectividade SSH
suid=7
sname=Script Assíncrono: 10
suser=
spid=17025
dst=192.168.0.2
dhost=marte
dport=22

Example 2

msg=Sessao iniciada para marte (192.168.0.1) - Usuario Privilegiado TPL - root pelo usuario Usuario Teste (usuario.teste)
suid=297
sname=Usuario Teste
suser=usuario.teste
spid=14410
dst=192.168.0.1
dpt=22
duser=root

I'm using https://regex101.com/ to help me

Any help is appreciated

[rahulrawlani]

Hello Friend,

If you want to extract the fields for both the events with one regex, then try using the below regex

.*msg\=(?.*)?suid\=(?\d+)?\ssname\=(?.*)?suser\=(?.*)?\sspid\=(?\d+)\sdst\=(?\d+\.\d+\.\d+\.\d+)\s(?:dhost\=(?\w+)\sdport\=(?\d+)|(?:dpt\=(?\d+)\sduser\=(?\w+)))

Let me know if that works for you.

[richgalloway · Accepted solution]

The examples are little different, which makes finding a common regex string tricky, but I believe this will work.

"msg=(?<msg>.*?) suid=(?<suid>\d+) sname=(?<sname>.*?) suser=(?<suser>.*) spid=(?<spid>\d+) dst=(?<dst>[^ ]+) (dhost=(?<dhost>.*?) dport=(?<dport>\d+))?(dpt=(?<dpt>\d+) duser=(?<duser>.*))?"

[kschmeling]

Perfect, it's working. The trick is this expression:

(dhost=(?<dhost>.*?) dport=(?<dport>\d+))?(dpt=(?<dpt>\d+) duser=(?<duser>.*))?

Thanks a lot!

[Richfez]

I am surprised this isn't happening automatically. Under normal circustances, Splunk would find any "key=value" pairs in inputs.

Have you repeated your search in Verbose mode to make sure you can see the fields it's already found?

[richgalloway]

I believe the default parser uses space as a delimiter so he'd end up with 'msg=O' and 'msg=Sessao'.

[kschmeling]

Yes, Splunk give me this results:

msg=O
msh=Sessao
sname=Script
sname=Usuario

[Richfez]

Yes, I realized that would have been the case after I had another cup of coffee and the wonderfully talented richgalloway reminded me of same. 🙂

There should be an answer below - give it a try and if it works for you, please accept it!

Happy Splunking!

-Rich