re.sub and named groups

[Emanuele D'Arrigo]

Hi everybody,

I'm having a ball with the power of regular expression but I stumbled
on something I don't quite understand:

theOriginalString = "spam?P<first>.*) ham?P<second>.*)"
aReplacementPattern = "\(\?P<first>.*\)"
aReplacementString= "foo"
re.sub(aReplacementPattern , aReplacementString, theOriginalString)

results in :

"spam:foo"

instead, I was expecting:

"spam:foo ham:"

Why is that?

Thanks for your help!

Manu

 

[MRAB]

> Hi everybody,
>
> I'm having a ball with the power of regular expression but I stumbled
> on something I don't quite understand:
>
> theOriginalString = "spam?P<first>.*) ham?P<second>.*)"
> aReplacementPattern = "\(\?P<first>.*\)"
> aReplacementString= "foo"
> re.sub(aReplacementPattern , aReplacementString, theOriginalString)
>
> results in :
>
> "spam:foo"
>
> instead, I was expecting:
>
> "spam:foo ham:"
>
> Why is that?
>
> Thanks for your help!
>

The quantifiers eg "*" are normally greedy; they try to match as much as
possible. Therefore ".*" matches:

spam?P<first>.*) ham?P<second>.*)
^^^^^^^^^^^^^^^^^^^^^

You could use the lazy form "*?" which tries to match as little as
possible, eg "\(\?P<first>.*?\)" where the ".*?" matches:

spam?P<first>.*) ham?P<second>.*)
^^

giving "spam:foo ham?P<second>.*)".

 

[Emanuele D'Arrigo]

You could use the lazy form "*?" which tries to match as little as
possible, eg "\(\?P<first>.*?\)" where the ".*?" matches:
spam?P<first>.*) ham?P<second>.*)
giving "spam:foo ham?P<second>.*)".

A-ha! Of course! That makes perfect sense! Thank you! Problem solved!

Ciao!

Manu

 

[Yapo Sébastien]

Hi everybody,

I'm having a ball with the power of regular expression but I stumbled
on something I don't quite understand:

theOriginalString = "spam?P<first>.*) ham?P<second>.*)"
aReplacementPattern = "\(\?P<first>.*\)"
aReplacementString= "foo"
re.sub(aReplacementPattern , aReplacementString, theOriginalString)

results in :

"spam:foo"

instead, I was expecting:

"spam:foo ham:"

Why is that?

Thanks for your help!

Manu

I think that .* in your replacement pattern matches ".*)
ham?P<second>.*" in your original string which seems correct for a regexp.
Perhaps you should try aReplacementPattern = "\(\?P<first>\.\*\)" or use
replace() since your replacement pattern is not a regexp anymore.

Sebastien

 

[Aahz]

I'm having a ball with the power of regular expression but I stumbled
on something I don't quite understand:

Book recommendation: _Mastering Regular Expressions_, Jeffrey Friedl

 

[Shawn Milochik]

Book recommendation: _Mastering Regular Expressions_, Jeffrey Friedl

I wholeheartedly second this! The third edition is out now.

 

[Paul McGuire]

Hi everybody,

I'm having a ball with the power of regular expression

Don't forget the ball you can have with the power of ordinary Python
strings, string methods, and string interpolation!

originalString = "spam:%(first)s ham:%(second)s"
print originalString % { "first" : "foo" , "second" : "" }

prints

spam:foo ham:

with far fewer surprises and false steps. (Note: Py3K supports this
same feature, but with different interpolation syntax, which I have
not learned yet.)

Book recommendation: Text Processing in Python, by David Mertz (free
online - http://gnosis.cx/TPiP/), in which David advises against
dragging out the RE heavy artillery until you've at least thought
about using ordinary string methods.

-- Paul

 

[Rhodri James]

Don't forget the ball you can have with the power of ordinary Python
strings, string methods, and string interpolation!

So the moral of this story is take a ball of strings with you for
when you get lost in regular expressions.