From: Robert Klemme <shortcutter@...>
Date: 2010-07-06T00:51:15+09:00
Subject: Re: File data extraction

2010/7/5 Brian Candler <b.candler@pobox.com>:
> Rolf Pedersen wrote:
>> The code now looks like this:
>>
>> filenames = File.open(filename).readlines.join.scan(/^@begin
>> Objects\n(.*?)^@end;/m)[0][0].split("\n").map{|l|
>> l.scan(/"(.*?)"/)[0][0]}
>> Probably far from optimal, but it seems to do the trick.
>
> That's the most important thing :-)
>
> I actually misread your example. If there's only one @begin Objects
> section, then 'scan' is overkill; a simple regexp match will do.
>
> res = if File.read(filename) =~ /^@begin Objects$(.*?)^@end;$/m
>  $1.scan(/^\s*"(.*?)"/).map { |r| r.first }
> end

If files are large than the line based approach is usually more
feasible.  In this case you can use the flip flop operator in an if
condition to select the lines we want:

17:31:49 Temp$ ./lextr.rb
["n_cst_xml_utils.sru", "n_melding.sru"]
17:48:32 Temp$ cat lextr.rb
#!/bin/env ruby19

ar = []

DATA.each_line do |line|
  if /^@begin Objects/ =~ line .. /^end;/ =~ line
    name = line[/^\s*"([^"]*)"/, 1] and ar << name
  end
end

p ar

__END__
Save Format v3.0(19990112)
@begin Libraries
 "felles.pbl" "";
@end;
@begin Objects
 "n_cst_xml_utils.sru" "felles.pbl";
 "n_melding.sru" "felles.pbl";
@end;
17:49:30 Temp$

Kind regards

robert


-- 
remember.guy do |as, often| as.you_can - without end
http://blog.rubybestpractices.com/