[lnkForumImage]
TotalShareware - Download Free Software
Confronta i prezzi di migliaia di prodotti.
Asp Forum
 Home | Login | Register | Search 


 

Forums >

comp.lang.ruby

Can't get subgroup of regex to repeat with +... what the ?

Jon Fi

5/16/2007 8:27:00 PM

I'm trying to match these kinds of malformatted xml tags. I'm beginning
to question my sanity, so i'm posting here.

Example Strings:
=====
A=" <orderMsg biz=0>"
B=" <orderMsg type=7 size=0>"
C=" <orderMsg type=7 size=0 biz=1>"
=====


I've come up with this regex:
=====
/<(\w+?)(?:\s(\w+)=(\w+))+>/
=====


But when matching string B from above:
=====
md=/<(\w+?)(?:\s(\w+)=(\w+))+>/.match(B)
=====


It will do this:
======
md[0]=<orderMsg type=7 size=0>
md[1]=orderMsg
md[2]=size
md[3]=0
nil
nil
nil
nil
=======


Why isn't the final + sign making the pattern "(?:\s(\w+)=(\w+))"
repeat?

As an exercise... /<(\w+?)(?:\s(\w+)=(\w+))(?:\s(\w+)=(\w+))>/ DOES
match String B from above. What the heck???

--
Posted via http://www.ruby-....
8 Answers

WoNáDo

5/16/2007 8:57:00 PM

0

Jon wrote:
> B=" <orderMsg type=7 size=0>"
> ...
> /<(\w+?)(?:\s(\w+)=(\w+))+>/
> ...
> md[0]=<orderMsg type=7 size=0>
> md[1]=orderMsg
> md[2]=size
> md[3]=0

It is correct. "(?:\s(\w+)=(\w+))+" matches two times, the last match is
with "size" and "0". The groups will be overwritten each time the "+"
will repeat the group.

Wolfgang Nádasi-Donner

--
Posted via http://www.ruby-....

Jon Fi

5/16/2007 8:59:00 PM

0

Wolfgang Nádasi-donner wrote:
> Jon wrote:
>> B=" <orderMsg type=7 size=0>"
>> ...
>> /<(\w+?)(?:\s(\w+)=(\w+))+>/
>> ...
>> md[0]=<orderMsg type=7 size=0>
>> md[1]=orderMsg
>> md[2]=size
>> md[3]=0
>
> It is correct. "(?:\s(\w+)=(\w+))+" matches two times, the last match is
> with "size" and "0". The groups will be overwritten each time the "+"
> will repeat the group.
>
> Wolfgang Nádasi-Donner

Ah ok. So how can I get it to repeat without overwriting the existing
values for the group? Or is there a better way to do this?

--
Posted via http://www.ruby-....

WoNáDo

5/16/2007 9:35:00 PM

0

Jon Fi wrote:
> Ah ok. So how can I get it to repeat without overwriting the existing
> values for the group? Or is there a better way to do this?

I would do it somehow like:

========== code ==========
texts = [ "<orderMsg biz=0>",
"<orderMsg type=7 size=0>",
"<orderMsg type=7 size=0 biz=1>"]

texts.each do |txt|
if (md=txt.match(/<(\w+?)((?:\s\w+=\w+)+)>/))
puts "\nkey '#{md[1]}' found"
md[2].scan(/\s(\w+)=(\w+)/) do |k, v|
puts " parameter '#{k}' has value '#{v}'"
end
else
puts "+++ no match for '#{txt}'"
end
end
========= result =========
key 'orderMsg' found
parameter 'biz' has value '0'

key 'orderMsg' found
parameter 'type' has value '7'
parameter 'size' has value '0'

key 'orderMsg' found
parameter 'type' has value '7'
parameter 'size' has value '0'
parameter 'biz' has value '1'
========== end ===========

Wolfgang Nádasi-Donner

--
Posted via http://www.ruby-....

Harry Kakueki

5/17/2007 1:38:00 AM

0

On 5/17/07, Jon <exabrial@gmail.com> wrote:
>
> Example Strings:
> =====
> A=" <orderMsg biz=0>"
> B=" <orderMsg type=7 size=0>"
> C=" <orderMsg type=7 size=0 biz=1>"
> =====
>
> I've come up with this regex:
> =====
> /<(\w+?)(?:\s(\w+)=(\w+))+>/
> =====
>
>
> But when matching string B from above:
> =====
> md=/<(\w+?)(?:\s(\w+)=(\w+))+>/.match(B)
> =====
>
>
> Why isn't the final + sign making the pattern "(?:\s(\w+)=(\w+))"
> repeat?
>
> As an exercise... /<(\w+?)(?:\s(\w+)=(\w+))(?:\s(\w+)=(\w+))>/ DOES
> match String B from above. What the heck???
>
> --
> Posted via http://www.ruby-....
>
>
Hi,

Unless you really want to write one regular expression for it all, you
could do something like this.

Split on spaces, then on '=' . Then process however you want.

r = B.strip.split(/\s/)
p r
r[1..-1].each {|f| p f.split("=")}

Harry


--

A Look into Japanese Ruby List in English
http://www.ka...

Harry Kakueki

5/17/2007 4:22:00 AM

0

On 5/17/07, Harry Kakueki <list.push@gmail.com> wrote:
> On 5/17/07, Jon <exabrial@gmail.com> wrote:
> >
> > Example Strings:
> > =====
> > A=" <orderMsg biz=0>"
> > B=" <orderMsg type=7 size=0>"
> > C=" <orderMsg type=7 size=0 biz=1>"
> > =====
>
> Unless you really want to write one regular expression for it all, you
> could do something like this.
>
> Split on spaces, then on '=' . Then process however you want.
>
> r = B.strip.split(/\s/)
> p r
> r[1..-1].each {|f| p f.split("=")}
>
> Harry
>

Sorry for the double post.
This is a little cleaner and easier, I think.

C.strip.delete("<>").split(/\s/).each {|f| p f.split("=")}

Harry


--

A Look into Japanese Ruby List in English
http://www.ka...

Robert Klemme

5/18/2007 7:55:00 AM

0

On 16.05.2007 22:59, Jon Fi wrote:
> Wolfgang Nádasi-donner wrote:
>> Jon wrote:
>>> B=" <orderMsg type=7 size=0>"
>>> ...
>>> /<(\w+?)(?:\s(\w+)=(\w+))+>/
>>> ...
>>> md[0]=<orderMsg type=7 size=0>
>>> md[1]=orderMsg
>>> md[2]=size
>>> md[3]=0
>> It is correct. "(?:\s(\w+)=(\w+))+" matches two times, the last match is
>> with "size" and "0". The groups will be overwritten each time the "+"
>> will repeat the group.
>>
>> Wolfgang Nádasi-Donner
>
> Ah ok. So how can I get it to repeat without overwriting the existing
> values for the group?

You can't.

> Or is there a better way to do this?

Probably. I am not sure what you are up to but you can use a two stage
approach like this:

texts = [
" <orderMsg biz=0>",
" <orderMsg type=7 size=0>",
" <orderMsg type=7 size=0 biz=1>",
]

texts.each do |t|
p t
md = /<([^\s>]+)((?:\s+\w+=\d+)*)/.match t

if md
tag = md[1]
attrs = md[2]

puts tag

attrs.scan(/(\w+)=(\d+)/) do |m|
print m[0], "=>", m[1], "\n"
end
end
end

Kind regards

robert

Harry Kakueki

5/18/2007 11:21:00 AM

0

On 5/17/07, Jon Fi <exabrial@gmail.com> wrote:
>
> Ah ok. So how can I get it to repeat without overwriting the existing
> values for the group? Or is there a better way to do this?
>

If you want to use regular expressions, try 'scan'.

c=" <orderMsg type=7 size=0 biz=1>"
c.scan(/\w+=?\w+/).each {|f| p f.split("=")}

Modify the regular expression as necessary.

Harry



--

A Look into Japanese Ruby List in English
http://www.ka...

Jon Fi

5/18/2007 4:11:00 PM

0

Harry Kakueki wrote:
> On 5/17/07, Jon Fi <exabrial@gmail.com> wrote:

> If you want to use regular expressions, try 'scan'.
>
> c=" <orderMsg type=7 size=0 biz=1>"
> c.scan(/\w+=?\w+/).each {|f| p f.split("=")}
>
> Modify the regular expression as necessary.
>
> Harry

Brilliant. Exactly what i was looking for.

--
Posted via http://www.ruby-....