Question about a regular expression

[Rob Hodges]

"Darrell" <news at dorb.com> writes:

> This doesn't work if you have nested parens.
> 
> >>> s='(1(2))'
> >>> print re.search(r"\((.*?)\)",s).groups()
> ('1(2',)
> >>>

True, but it works fine for what Yoav wanted to do, which was to grab
the tokens out of parens that weren't nested (at least in the examples
offered) --

>>> s='a(x,y)b(x)'
>>> for st in re.findall(r"\((.*?)\)", s):
...     print string.split(st, ",")
... 
['x', 'y']
['x']

If they were nested, we'd need to know more precisely what defines a
token to approach the problem.  For example if it were
'a(w,x(y))b(x)', are 'w' and 'x(y)' the tokens belonging to a, and in
turn 'y' the token belonging to x -- or are 'w', 'x' and 'y' what we
want to extract for a?  It depends on the meaning of the data and the
detail you want to extract... and frankly, I ain't *touchin'* that
unless I have to ;)

Merry Christmas, Pythonistas!

-Rob

> > > Yoav I H Parish <parish at ikb.mavt.ethz.ch> writes:
> > >
> > > > i have a string which could look something like
> > > >
> > > >     a(x,y)b(x)
> > > >     or
> > > >     c(x,y,z)b(x)a(x,y)