From: MonkeeSage <MonkeeSage@...>
Date: 2007-12-06T09:10:01+09:00
Subject: Re: hash

On Dec 5, 4:49 pm, Johnathan Smith <stu...@hotmail.com> wrote:
> hey guys
>
> sorry for the whole mix up! I appreciate all the help I've received
> especially as im new to ruby! im also new to the forum and should have
> tried to stick it our with the first threat! so i hope i havn't offended
> any of you by this.

Nah! Any mistake you might make, we've already done that and worse
(probably several times). :)

> that aside, i really hope you can help me resolve this problem.
>
> Jordan, the code you gave me worked successfully. However, it just deals
> with the the two types i gave you, author and type.Its partially my
> fault for not explaining properly becuase i hope it could deal with any
> information after the tag. is this possible to achieve?

Sure it is. The code I gave you can be expanded, but it was mainly
just for you to get started with, because it is fragile (as you saw
from my mis-post). Even with the correct "author[8..-1]", if you
accidentally leave out the space between the "Author: Name" line, it
will still clip the first letter off the name ([number..number] means
to grab a smaller section from a larger container, and -1 has the
special meaning of the last item; so author[8..-1] means to get the
characters from number 8 to the end).

What you really want is something more flexible, that will allow for
those kinds of mistakes and still do the right thing in almost every
case. That is a regular expression. It lets you give it a pattern
which has certain rules, and matches a string in different ways based
on those rules. And this is just what Phrogz' code is doing:

# we'll get rid of linecount since you can
# get that information directly from your hash

# and we'll store the hashes as subhashes
# underneath a main one, keyed by "Tag:" value

database = {}

File.open('reference.txt') { | handle |

  last_tag = nil # we'll use this to keep track of
                 # the last tag we've seen, see below

  handle.each { | line |

    # the following regexp says to match:
    #   ^          = starting at the first char
    #   (\w+)      = 1 or more "word" characters (a-z / 0-9)
    #   :          = followed by a literal colon
    #   \s*        = then zero or more "space" characters
    #  ([\w+,\s]+) = 1 or more "word/space/," characters
    #  $           = with nothing else until the end of line
    #
    # any string meeting that criteria is a match and m is
    # the match data, however if the string doesn't match
    # e.g., a blank line, m is nil

    m = line.match(/^(\w+):\s*([\w+,\s]+)$/)

    if m # if m is a match (i.e., not nil)

      if m[1] == 'Tag' # we want to add a key to the hash
                       # since this is a "Tag:" line
                       # when we used the () in the regexp
                       # that told ruby we only care about
                       # those chars, and we can access
                       # them in the match object with m[n]
                       # i.e, m[1] means whatever part of the
                       # line matched (\w+)

        last_tag = m[2].chomp
        database[last_tag] = {} # make a subhash as the value
                                # now we have, e.g.,
                                # database["ref1"] = {}
      else

        # it's not a "Tag:" line, so we add it as key, value
        # to the subhash we created for the last tag key we
        # created in the hash

        database[last_tag][m[1].downcase] = m[2].chomp
        # e.g., database["ref1"]["type"] = "Journal"

      end
    end
  }
}

# now print it
print "# of Tags: ", database.length, "\n\n"
print "Contents: ", database.inspect, "\n\n"

puts database["ref2"]["author"]


========

This outputs:

# of Tags: 3

Contents: {"ref1"=>{"author"=>"Little, S R", "type"=>"Book"},
"ref2"=>{"author"=>"Smith, J", "type"=>"Journal"},
"ref3"=>{"author"=>"Williams, M", "type"=>"Conference Paper"}}

Smith, J

> again sorry for the mix up
> cheers
>
> --
> Posted viahttp://www.ruby-forum.com/.

No worries :)

HTH,
Jordan