Shell Programming and Scripting

Hi all Unix Gurus!

Since hours (even days :-)) I'm trying to find the correct pattern to search for IP addesses in text files.
The pattern to find a IP address itself is not too difficult:
'(([0-9]|[1-9][0-9]|1[0-9]{2}|2[0-4][0-9]|2[0-5]{2})\.){3,}([0-9]|[1-9][0-9]|1[0-9]{2}|2[0-4][0-9]|2[0-5]{2})'

BUT, of course the above pattern is also matching lines like
v.55.25.1.7v
1.3.6.1.4.1.897.4.6.1 = dce
394983|12.50.1.0.0|
--> which are not really IP addresses

The big question is now to avoid the matching of above lines. I thought the best is to NOT ALLOW A DOT [^.] BEFORE AND AFTER MY PATTERN:
'[^.](([0-9]|[1-9][0-9]|1[0-9]{2}|2[0-4][0-9]|2[0-5]{2})\.){3,}([0-9]|[1-9][0-9]|1[0-9]{2}|2[0-4][0-9]|2[0-5]{2})[^.]'

BUT THIS IS SIMPLY NOT WORKING ! It is still showing above example lines

I'm using nawk and egrep on Solaris 9.

Many, many thanks in advance for any hint on my problem.

The masq I use is this one:

Code:

masq='^((25[0-5]|2[0-4][0-9]|1?[0-9]?[0-9])\.){3}(25[0-5]|2[0-4][0-9]|1?[0-9]?[0-9])$'
 # to test:

$ echo "ip" | egrep "$masq"

Hi ripat,

thanks for your reply, but this is unforunately not working as you asume that the IP address is the only information in the line (^....$).

Here is my testfile:

Code:

1. Yes 123.25.67.8 first line
2. RIP1:171.27.38.166:spooler:spooler:/HERMES:bin
3. this is an IP128.26.18.7:
4. 81.1.0.0 valid
5. 123.114.45.6c valid
6. valid IP |12.50.1.0|
7. this is also not an IP4.0.1.1.C
8. v.55.25.1.7vNO IP
9. 20040908|394983|12.50.1.0.0|SZ/WIS/SEITE01||
10. 1.3.6.1.4.1.897.4.6.1   = dce

--> I expect line 1-6 to be found, but not 7-10

As RFC is not my favorite reading, can you define what you would consider as valid IP? The regex I gave, covers bytes going from 0 to 255.

Edit:

The following pattern works on your sample file but I am sure this is getting too specific.

Code:

'[^.0-9]((25[0-5]|2[0-4][0-9]|1?[0-9]?[0-9])\.){3}(25[0-5]|2[0-4][0-9]|1?[0-9]?[0-9])[^\.]'

Hi ripat,

THANKS A LOT! Nearly there :-))

I consider the whole range of IPs as valid IP (0-255.0-255.0-255.0-255).

The only problem left (which I can see currently) is that your regexp is also ignoring following additional line in my testfile (which is valid IP):

Code:

11. 12.50.1.0

Just add the start and end of string anchors at the begining and at the end of the pattern:

Code:

'(^|[^.0-9])((25[0-5]|2[0-4][0-9]|1?[0-9]?[0-9])\.){3}(25[0-5]|2[0-4][0-9]|1?[0-9]?[0-9])([^\.]|$)'

Hi ripat,

BRILLIANT! Thanks so much!

Your pattern is working (I guess you missed the NOT^ at the beginning of [.0-9]: [^.0-9] instead of [.0-9]):

Code:

'(^|[^.0-9])((25[0-5]|2[0-4][0-9]|1?[0-9]?[0-9])\.){3}(25[0-5]|2[0-4][0-9]|1?[0-9]?[0-9])([^\.]|$)'

IT IS WORKING FINE, BUT I DO NOT UNDERSTAND WHY


(^|[^.0-9]) = "beginning of the line" OR "not dot, not number"

((25[0-5]|2[0-4][0-9]|1?[0-9]?[0-9])\.){3}(25[0-5]|2[0-4][0-9]|1?[0-9]?[0-9])= valid IP (range 0-255.0-255.0-255.0-255)

([^\.]|$) = "not backslash, not dot" OR "end of the line"

Am I interpreting the regexp correctly?