From: Florian Gross <flgr@...>
Date: 2004-05-29T10:58:40+09:00
Subject: Re: regular expression help please

Paul wrote:

> How do I extract the name and value from the following lines:
> 
> name=paul value=10 otherstuff=123
> 
> but the line may also be:
> name='hello paul' value='10' otherstuff='123'

irb(main):001:0> lines.scan(/^name=('?)(.*?)\1\s+value=('?)(\S*?)\3/)
=> [["", "paul", "", ""], ["'", "hello paul", "'", "10"]]

Things start to get more interesting when Strings can also contain 
quoted delimiters however. (As in 'Don\'t use PHP!')

Regexp::English lets us solve that case relatively easily however:

> irb(main):035:0> re = Regexp::English.new do
> irb(main):036:1*   quoted_string = quoted_text("'")
> irb(main):037:1>   unquoted_string = non_whitespace
> irb(main):038:1>   name_val = (quoted_string | unquoted_string).capture(:name)
> irb(main):039:1>   value_val = (quoted_string | unquoted_string).capture(:value)
> irb(main):040:1>   literal("name=") + name_val + whitespace +
> irb(main):041:1*   literal("value=") + value_val
> irb(main):042:1> end
> => /name=((?x:'((?x:(?!\\).(?:\\{2})?\\'|(?!').)*)'|\S+))\s+value=((?x:'((?x:(?!\\).(?:\\{2})?\\'|(?!').)*)'|\S+))/
> irb(main):051:0> lines = %{
> irb(main):052:0"   name='hello. I\\'m paul' value='don\\'t do that'
> irb(main):053:0"   name=foobar value=3
> irb(main):054:0"   name='drei' value='three'
> irb(main):055:0" }
> irb(main):070:0> lines.scan(re)
> => [["'hello. I\\'m paul'", "hello. I\\'m paul", "'don\\'t do that'", "don\\'t do that"],
>     ["foobar",  nil, "3", nil],
>     ["'drei'", "drei", "'three'", "three"]]

Regards,
Florian Gross