2

I'm attempting to parse a series of SHOW CDP NEIGHBORS DETAIL outputs so I can capture each device and its ip address.

The issue that I am coming across is that some devices may have multiple ip addresses assigned to it, here is an example output.

Device ID: RTPER1.MFN21Mb.domain.local
Entry address(es): 
  IP address: 200.152.51.3
  IP address: 82.159.177.233
  IP address: 201.152.51.140
  IP address: 84.252.100.3
Platform: Cisco 2821,  Capabilities: Router Switch IGMP

I wrote this regex to capture the input, and according to gskinner it matches all 4 ip addresses, but the capture is just the last one (as expected from regex)

Device ID: ([0-9A-Za-z-.&]+)\s+Entry address\(es\):\s+(?:IP address: ([0-9.]+)\s+)+

So I went online to figure out how to do this. I tried teh regex suggested here Capturing repeating subpatterns in Python regex but using the regex module did not change the output. I still only get the last ip address on the list, and none of the others.

Following the example I get

temp = regex.match(r'Device ID: ([0-9A-Za-z-.&]+)\s+Entry address\(es\):\s+(?:IP address: ([0-9.]+)\s+)+', file)
print temp

Temp returns None.

If I do findall. I get a return of just the last ip address 84.252.100.3

If I add multiple capture groups, such as 

temp = regex.findall(r'Device ID: ([0-9A-Za-z-.&]+)\s+Entry address\(es\):\s+(?:IP address: ([0-9.]+)\s+)?\s+(?:IP address: ([0-9.]+)\s+)?\s+(?:IP address: ([0-9.]+)\s+)?\s+(?:IP address: ([0-9.]+)\s+)?\s+(?:IP address: ([0-9.]+)\s+)?', file)
print temp

It only matches the ones that have mutliple ip addresses, and not the others

Hopefully someone can help.

Community
  • 1
  • 1
iargue
  • 69
  • 2
  • 13

1 Answers1

1

As far as I'm aware, only .NET allows you to iterate through quantified (repeated) capturing groups. Consider this (finite) alternative:

Device ID: ([0-9A-Za-z-.&]+)\s+Entry address\(es\):\s+(?:IP address: ([0-9.]+)\s+)(?:IP address: ([0-9.]+)\s+)?(?:IP address: ([0-9.]+)\s+)?(?:IP address: ([0-9.]+)\s+)?
                                                                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

This will capture up 1 IP address in $2 and up to three more in $3, $4, and $5. (I'm using the $ notation idiomatically, of course.) You can add as many as you want. If you need all of the IP addresses to be present in a single group, i.e. $2, then your only choice is to include the text with them:

Device ID: ([0-9A-Za-z-.&]+)\s+Entry address\(es\):\s+((?:IP address: (?:[0-9.]+)\s+)+)
                                                      ^                ^^             ^
Andrew Cheong
  • 29,362
  • 15
  • 90
  • 145
  • 1
    I haven't used it myself, but the Python `regex` module seems to allow getting a list of text captured by an group that is repeated, as per the answer in the question that OP links to. (Anyway, the first option is not even an option, IMO, since you are imposing an upper limit) – nhahtdh Apr 09 '13 at 17:19
  • 1
    @nhahtdh - Ah, my mistake. I only read the first answer, and I'm not familiar with Python myself. I recognize your name from past regex discussions. So with due respect, I don't believe that imposing an upper limit makes an option invalid. Especially under the (now found to be incorrect) premise that there's no other way, being able to specify a regex that can capture up to, say, 100 IP addresses, may be more than sufficient for OP's practical needs. – Andrew Cheong Apr 09 '13 at 18:45
  • I think it is more practical to use the second option in your post (take the whole block of homogenous data and use a second pass to extract it). My intention is to discourage the use of the first approach, unless it is a one-off business. – nhahtdh Apr 09 '13 at 19:24
  • I'll accept this answer as the solution because it tends to answer what I wanted to know. It is sad that python doesn't support repeated capture groups, but luckily there is a work around. As with nhahtdh suggestion, I went with taking the entire block and then parsing it a second time, that way if some servers have more then 5 ip addresses I still get them all. thanks for the help. – iargue Apr 10 '13 at 15:42