ansaurus

Question

Extract a specific portion of connection string

Answer 1

+1 A:

You can use a couple of string.split calls, the first using '@' as a separator, the second using '.'

Ofir 2010-07-27 07:45:54

+1 i think it's really the best way. Neither regex solutions given matche string without @.

M42 2010-07-27 09:12:11

Answer 2

A:

do a split by '@', and then substring.

Ankit Jain 2010-07-27 07:46:00

Answer 3

+1 A:

>>> x = "[email protected]"
>>> x.split("@")[1].split(".")[0]
'host'
>>> y = "first.last@host"
>>> y.split("@")[1].split(".")[0]
'host'
>>>

There will be an IndexError Exception thrown if there is no @ in the string.

Fabian 2010-07-27 07:48:54

Answer 4

A:

'[email protected]'.split('@')[1].split('.')[0]

CrociDB 2010-07-27 07:49:39

Answer 5

+1 A:

host = re.search(r"@(\w+)(\.|$)", s).group(1)

Amarghosh 2010-07-27 07:54:58

Doesn't look like it'll work if there's no `.` after the `@`

John Machin 2010-07-27 08:00:51

@John You're right - fixed it.

Amarghosh 2010-07-27 08:04:33

Nice one. I added a `|` after the `@` in case there is no `@`

Alix 2010-07-27 08:10:48

@Amarghosh: You didn't fix it properly; `$` can match `\n` at the end of the string; use `\Z` instead of `$`.

John Machin 2010-07-27 08:18:41

@Alix: Why? Apart from the `\Z` thing, the pattern fails to match in other circumstances e.g. `foo@#` so just leave failure to match as being generally "invalid input".

John Machin 2010-07-27 08:23:00

@John That's a feature I need: to be able to specify a server name only e.g. `foo` or `foo.bar.tld` (it was not in my question though).

Alix 2010-07-27 08:29:57

@John what if the string contains multiple lines and a line ends with host name as in `"user@host\nsome other text"`

Amarghosh 2010-07-27 09:03:27

Will not matche the first 2 examples from OP: ie without @ in the string.

M42 2010-07-27 09:09:23

Yup, I'm working on it. Can't get it working without a `if:` `else:` though...

Alix 2010-07-27 09:12:53

@Alix, @Amarghosh: not finished with you yet :) ... using `\w` allows the underscore `_` (dodgy) but disallows the hyphen `-` (definitely not a good idea) ... see http://en.wikipedia.org/wiki/Hostname ... I suggest use of `[-\w]` or `[-A-Za-z0-9]` instead of `\w`

John Machin 2010-07-27 09:14:54

@Amarghosh: If the OP wants to allow trailing rubbish (including a newline), then just use `@([-\w]+)` with no end anchor.

John Machin 2010-07-27 09:23:44

@John `[-A-Za-z0-9]` is the correct one.

Amarghosh 2010-07-27 10:01:51

@Amarghosh: I know that [-A-Za-z0-9] is the "correct" one, but out there in real world some folks have host names with underscores in them ...

John Machin 2010-07-27 10:47:55

Answer 6

A:

Here is one more solution:

re.search("^.*@([^.]*).*", str).group(1)

edit: Much better solution thanks to the comment:

re.search("@([^.]*)[.]?", str).group(1)

Klark 2010-07-27 07:55:56

(1) `re.search("^.*X")` is functionally equivalent to `re.search("X")` [except in the presence of newlines] and (as implemented) is horribly slower if X is not found. (2) The trailing `.*` (matches 0 or more of any character [except a newline]) is utterly pointless.

John Machin 2010-07-27 08:11:40

Thanks for the comment. Those are the great things to know.

Klark 2010-07-27 08:18:43

Any trailing pattern that can match 0 characters is utterly pointless, whether it's `.*` or `[.]?` or something else -- you don't want what it matches, so don't put it in.

John Machin 2010-07-27 09:55:43

Answer 7

A:

>>> s="[email protected]"
>>> s[s.index("@")+1:]
'host.domain.com'
>>> s[s.index("@")+1:].split(".")[0]
'host'

ghostdog74 2010-07-27 08:00:03

Answer 8

A:

import re

hosts = """
user@host1
first.last@host2
[email protected]
[email protected]
"""

print re.findall(r"@(\w+)", hosts)

returns:

['host1', 'host2', 'host3', 'host4']

ssoler 2010-07-27 08:05:36

ansaurus

tags:

views:

answers:

Extract a specific portion of connection string

related questions