regex problem

[Odd-R.]

Input is a string of four digit sequences, possibly
separated by a -, for instance like this

"1234,2222-8888,4567,"

My regular expression is like this:

rx1=re.compile( r"""\A(\b\d\d\d \d,|\b\d\d\d\d-\d\d\d\d,)*\Z"" ")

When running rx1.findall("12 34,2222-8888,4567,")

I only get the last match as the result. Isn't
findall suppose to return all the matches?

Thanks in advance.


--
Har du et kjøleskap, har du en TV
så har du alt du trenger for å leve

-Jokke & Valentinerne

[Thomas Guettler]

Re: regex problem

Am Tue, 26 Jul 2005 09:57:23 +0000 schrieb Odd-R.:
[color=blue]
> Input is a string of four digit sequences, possibly
> separated by a -, for instance like this
>
> "1234,2222-8888,4567,"
>
> My regular expression is like this:
>
> rx1=re.compile( r"""\A(\b\d\d\d \d,|\b\d\d\d\d-\d\d\d\d,)*\Z"" ")[/color]

Hi,

try it without \A and \Z

import re
rx1=re.compile( r"""(\b\d\d\d\d ,|\b\d\d\d\d-\d\d\d\d,)""")
print rx1.findall("12 34,2222-8888,4567,")
# --> ['1234,', '2222-8888,', '4567,']

Thomas

--
Thomas Güttler, http://www.thomas-guettler.de/

[John Machin]

Re: regex problem

Odd-R. wrote:[color=blue]
> Input is a string of four digit sequences, possibly
> separated by a -, for instance like this
>
> "1234,2222-8888,4567,"
>
> My regular expression is like this:
>
> rx1=re.compile( r"""\A(\b\d\d\d \d,|\b\d\d\d\d-\d\d\d\d,)*\Z"" ")
>
> When running rx1.findall("12 34,2222-8888,4567,")
>
> I only get the last match as the result. Isn't
> findall suppose to return all the matches?[/color]

For a start, an expression that starts with \A and ends with \Z will
match the whole string (or not match at all). You have only one match.

Secondly, as you have a group in your expression, findall returns what
the group matches. Your expression matches zero or more of what your
group matches, provided there is nothing else at the start/end of the
string. The "zero or more" makes the re engine waltz about a bit; when
the music stopped, the group was matching "4567,".

Thirdly, findall should be thought of as merely a wrapper around a loop
using the search method -- it finds all non-overlapping matches of a
pattern. So the clue to get from this is that you need a really simple
pattern, like the following. You *don't* have to write an expression
that does the looping.

So here's the mean lean no-flab version -- you don't even need the
parentheses (sorry, Thomas).
[color=blue][color=green][color=darkred]
>>> rx1=re.compile( r"""\b\d\d\d\d, |\b\d\d\d\d-\d\d\d\d,""")
>>> rx1.findall("12 34,2222-8888,4567,")[/color][/color][/color]
['1234,', '2222-8888,', '4567,']

HTH,
John

[Duncan Booth]

Re: regex problem

John Machin wrote:
[color=blue]
> So here's the mean lean no-flab version -- you don't even need the
> parentheses (sorry, Thomas).
>[color=green][color=darkred]
> >>> rx1=re.compile( r"""\b\d\d\d\d, |\b\d\d\d\d-\d\d\d\d,""")
> >>> rx1.findall("12 34,2222-8888,4567,")[/color][/color]
> ['1234,', '2222-8888,', '4567,'][/color]

No flab? What about all that repetition of \d? A less flabby version:
[color=blue][color=green][color=darkred]
>>> rx1=re.compile( r"""\b\d{4}( ?:-\d{4})?,""")
>>> rx1.findall("12 34,2222-8888,4567,")[/color][/color][/color]
['1234,', '2222-8888,', '4567,']

[John Machin]

Re: regex problem

Duncan Booth wrote:[color=blue]
> John Machin wrote:
>
>[color=green]
>>So here's the mean lean no-flab version -- you don't even need the
>>parentheses (sorry, Thomas).
>>
>>[color=darkred]
>>>>>rx1=re.com pile(r"""\b\d\d \d\d,|\b\d\d\d\ d-\d\d\d\d,""")
>>>>>rx1.findal l("1234,2222-8888,4567,")[/color]
>>
>>['1234,', '2222-8888,', '4567,'][/color]
>
>
> No flab? What about all that repetition of \d? A less flabby version:
>
>[color=green][color=darkred]
>>>>rx1=re.comp ile(r"""\b\d{4} (?:-\d{4})?,""")
>>>>rx1.findall ("1234,2222-8888,4567,")[/color][/color]
>
> ['1234,', '2222-8888,', '4567,']
>[/color]


OK, good idea to factor out the prefix and follow it by optional -1234.
However optimising re engines do common prefix factoring, *and* they
rewrite stuff like x{4} as xxxx.

Cheers,
John

[Odd-R.]

Re: regex problem

On 2005-07-26, Duncan Booth <duncan.booth@i nvalid.invalid> wrote:[color=blue][color=green][color=darkred]
>>>> rx1=re.compile( r"""\b\d{4}( ?:-\d{4})?,""")
>>>> rx1.findall("12 34,2222-8888,4567,")[/color][/color]
> ['1234,', '2222-8888,', '4567,'][/color]

Thanks all for good advice. However this last expression
also matches the first four digits when the input is more
than four digits. To resolve this problem, I first do a
match of this,

regex=re.compil e(r"""\A(\b\d{4 },|\d{4}-\d{4},)*(\b\d{4 }|\d{4}-\d{4})\Z""")

If this turns out ok, I do a find all with your expression, and then I get
the desired result.


--
Har du et kjøleskap, har du en TV
så har du alt du trenger for å leve

-Jokke & Valentinerne