From: Axel Etzold <AEtzold@...>
Date: 2007-08-10T18:22:00+09:00
Subject: Re: converting xml to MYSQL that reflects the "tag's embedding"?

> Dear Ruby guru's,
> 
> Can any of you tell me how to convert xml to SQL?
> 
> The tags of the text I need to convert are linguistic in nature. E.g.
> <sentence><clause><word><morpheme></morpheme></word></clause></
> sentence> and the like.
> 
> What I need, is a database that somehow reflects the embedding. Can
> this be done?
> 
> 
> Kind regards,
> Robert.
> 

Dear Robert,

I think what you need is to identify "brackets of text" starting with
something like <sentence> and ending with </sentence> etc., then pass
these brackets to some variable in Mysql.

For Ruby and MySql, see here: http://www.kitebird.com/articles/ruby-mysql.html

The finding of the brackets can be done using regular expressions.
I've written some code below.
You'd still need to choose some convenient variable name for the different
types of brackets (sentence,word etc...)

Best regards,

Axel

----------------------------------------------------------------

class String
	def find_positions(opening,closing)
		open_p=[]
		closing_p=[]
		temp=self.dup
		while opening.match(temp)
			ref=opening.match(temp)
			open_p<<temp.index(ref[0])
			temp.sub!(ref[0],' '*ref[0].length)
		end
		temp=self.dup
		while closing.match(temp)
			ref=closing.match(temp)
			closing_p<<temp.index(ref[0])
			temp.sub!(ref[0],' '*ref[0].length)
		end
		
		# correctly associate opening and closing "brackets"
		# check again, especially for nested brackets
		
		brackets=[]
		closing_p.each{|clb|
			temp=open_p.dup
			opb=(temp.delete_if{|y| y>clb}).max
			open_p=open_p-[opb]
			brackets<<[opb,clb]
		}
		return brackets.sort{|x,y|  x[0]<=>y[0]}
	end
	def find_brackets_text(opening,closing)
		pos_array=self.find_positions(opening,closing)
		text=[]
		p 'positions of brackets in the text'
		p pos_array
		pos_array.each{|op,cl|
		        p 'opcl'
			p op
			p cl
			text<<self[op...cl]
		}
		return text
	end
end
a="<s><c><w><s><m></m></s></w></c><s></s>blabla</s><s>bla</s>"
c=a.find_brackets_text(/<s>/,/<\/s>/)
p c



-- 
GMX FreeMail: 1 GB Postfach, 5 E-Mail-Adressen, 10 Free SMS.
Alle Infos und kostenlose Anmeldung: http://www.gmx.net/de/go/freemail