If I wanted to grep through a file or multiple at once....

DustinB3403

Sample output I'm trying to find would be

Windows Server 2008 R2 | SVR12.localdomain

If that helps, I know it's multiple expressions, but the formatting isn't working (I'm sure I broke it).

manxam

Wouldn't the following get every line containing a variation of "server" within this file?
grep -rin server server.txt > out.txt

DustinB3403

@manxam while that would, it would also get anything with the word Server listed, I need to find anything with the word Server and then some server-name in it, ideally piping only those details to a output file.

DustinB3403

I'm trying to break out a datadump essentially into unique records pertaining to "Servers" only and I'm not sure if I can get a generic enough regular expression to do just that.

DustinB3403

Is it possible to grep for a set number of characters around a regular expression and give me that? For example

"Server 2012 R2" so give me the previous 20 characters, Server 2012 R2, and the next 2012 characters?

manxam

@DustinB3403 said in If I wanted to grep through a file or multiple at once....:

Windows Server 2008 R2 | SVR12.localdomain

Something like
/\werver.+/g would get you "Server 2008 R2 | SVR12.localdomain"
You'd have to do a negative lookahead in order to capture the prior input. the /g gives you global results so that it doesn't stop at the first match.

DustinB3403

Okay I think I have a process that should work.

grep -E "{0.30}Windows Server.{0.30}" server.txt | sort --unique

Now just to get the remainder of the lines/next lines if they are wrapped.

manxam

@DustinB3403 : Well that looks like fun

DustinB3403

@manxam said in If I wanted to grep through a file or multiple at once....:

@DustinB3403 : Well that looks like fun

I assume you have a better approach? How would your approach look like, I just tested it and got no output.

manxam

@DustinB3403 : No, I definitely don't have a better approach, especially when you have to capture wrapped lines.
I'd just toss a sample of your input file into a regex tester online and build out your regex from there...

DustinB3403

@manxam said in If I wanted to grep through a file or multiple at once....:

@DustinB3403 : No, I definitely don't have a better approach, especially when you have to capture wrapped lines.
I'd just toss a sample of your input file into a regex tester online and build out your regex from there...

Doh I thought you may have had an idea. No worries, I'll keep at it.

DustinB3403

This is a small subset of the data I'm working with (anonymous) 

    On for All programs and services except those I select
OS Manufacturer:
Microsoft Corporation
OS Version:
5.2.3790 Service Pack 2 (Build 3790)
OS Caption:
Microsoft(R) Windows(R) Server 2003 Standard x64 Edition
OS Virtual Memory:
2528 MB
OS System Directory:
C:\WINDOWS\system32
OS Windows Directory:
C:\WINDOWS
OS Install Date:
8/5/2008 12:49:17 PM
DB-Server
Windows Server 2012 R2 Standard


Remote Listening Ports:
RDP (3389/TCP)DB-Virtual
Windows Server 2008 Standard

Now if I wanted to find "Windows Server 2012" and then the line above it, which is the server name. How in the heck would I do that?

Paging @scottalanmiller

DustinB3403

This will output the Server detail, but doesn't jump to the prior or next lines.

grep -riE "((.*\Windows Server){5}}*Windows Server" source.txt

manxam

Just spitballing again, but what about grep with -A# and -B# and a regex of .+?(Server.+)
A# stands for n lines "after" the match.
B# stands for m lines "before" the match.

Using grep -rni -E '.+?(Server.+)' -B1 should capture

DB-Server
Windows Server 2012 R2 Standard

DustinB3403

@manxam said in If I wanted to grep through a file or multiple at once....:

Just spitballing again, but what about grep with -A# and -B# and a regex of .+?(Server.+)
A# stands for n lines "after" the match.
B# stands for m lines "before" the match.

Using grep -rni -E '.+?(Server.+)' -B1 should capture
DB-Server
Windows Server 2012 R2 Standard

That I can at least work with, with relative ease. Still not perfect, but way better than the full details I was working with.

DustinB3403

This is the final working regex that was used for anyone else who may ever need this.

grep -ri -E '.+?(Windows Server.+)' -B7 source.txt > regex.txt

From

@manxam said in If I wanted to grep through a file or multiple at once....:

Just spitballing again, but what about grep with -A# and -B# and a regex of .+?(Server.+)
A# stands for n lines "after" the match.
B# stands for m lines "before" the match.

Using grep -rni -E '.+?(Server.+)' -B1 should capture
DB-Server
Windows Server 2012 R2 Standard

DustinB3403

Thanks a ton @manxam!

manxam

@DustinB3403 : Team effort!