From: Mark Hubbart <discordantus@...>
Date: 2005-06-23T11:47:19+09:00
Subject: Re: Getting a list of results from one regular expression

On 6/22/05, tietyt@gmail.com <tietyt@gmail.com> wrote:
> Hello I'm new to Ruby.  I've read most of the pragmatic programmer
> guide but couldn't find anything that explained how to do this.
> 
> To summarize my whole question: how do I get EVERY match of a regular
> expression (instead of just the first)?
> 
> Here's my situation, I've got this long string that contains XML.  I
> would like to parse it.  Specifically, I want to search this string for
> all instances of a pattern like /stringAlias="(.*)"/
> 
> I'm no pro with regex, but I think that will find a match for a string
> that looks like this: stringAlias="BLAH"
> 
> And because of the (.*), the result will be BLAH
> 
> Now this is all fine and good.  But what I can't figure out is how to
> get every match in an array (instead of just the first match.
> 
> If i have stringAlias="BLAH" ... stringAlias="BLEH" how do I get an
> array that is ["BLAH", "BLEH"]?
> 
> Keep in mind that there are a dynamic number of matches for
> stringAlias="(.*)"
> 
> 
> This is the code I wrote to try to do it:
> 
> def ...
>  @aliases = []
>  matchedData = /stringAlias="(.*?)"/.match(@data)
>  @aliases = matchedData.to_a
>  puts @aliases
> end
> 
> The length of the array is 2 and the result is this:
> stringAlias="OP"
> OP
> 
> Even though the data is this:
> <string RSLDefined="false" active="false" languageId="1"
>     sortOrder="0" stringAlias="OP">
>     <stringValue><![CDATA[Open or Pending]]></stringValue>
> </string>
> <string RSLDefined="false" active="true" languageId="1"
>     sortOrder="1" stringAlias="1">
>     <stringValue><![CDATA[Open]]></stringValue>
> </string>
> <string RSLDefined="false" active="true" languageId="1"
>     sortOrder="2" stringAlias="2">
>     <stringValue><![CDATA[Pend]]></stringValue>
> </string>
> <string RSLDefined="false" active="true" languageId="1"
>     sortOrder="3" stringAlias="3">
>     <stringValue><![CDATA[Decline]]></stringValue>
> </string>
> <string RSLDefined="false" active="true" languageId="1"
>     sortOrder="4" stringAlias="4">
>     <stringValue><![CDATA[Complete]]></stringValue>
> </string>

Regexp#match only gives the first match; the matchdata object is sort
of an array of the entire match, followed by the subexpression
matches. What you want is String#scan: (warning, untested)

  regexp = /stringAlias="(.*?)"/
  matches = @data.scan(regexp)

Since the regexp has a subexpression matcher, that is what will be put
into the array "matches". You'll get an array something like this:

  [["OP"],["1"],["2"], ... ]

(each match has it's own subarray, since it's a subexpression match)

Check out the docs for String#scan for more info...

cheers,
Mark