From: roktas <roktas@...>
Date: 2002-02-24T09:32:59+09:00
Subject: Re: Newbie question

Hi,

David and Alan (Chen), thanks for your quick replies. Let me first
clarify a point that this is not a real-world problem for me (otherwise
David's post would be enough.) I just want to tinker with some concepts
and also to realize an elegantly simple, effective solution in terms of
OOP principles.

> 
> What you want, I think, is an array of hashes.  Each hash would
> represent one line from your data file.
> 

Actually, this is the first approach I tried. But, with keys of hash
object, there would be 2*n records for n (equal) lines. So, I've come to
the following point (please do not care for lexical errors)

$FMT = "/\t/"

# Only one hash for the field labels (not 2*n)
$labels = {}

def parse_line(line, fmt=$FMT)
    # No extra process for the present
    line.chomp.strip(delim)
end

def DB
    def initialize(f)
        # First create the $labels hash from the first line
        $labels = ...
        
        # Other stuff filling the @tab array 
        ...
    end
    
    ...
    
    def [](key)
        @tab[key]
    end
end

def Record < Array
    def initialize(line)
        # Parse and create a new record
        parse_line(line)
    end
    def [](key)
        key if key.kind_of?Integer
        $labels.index(key) if key.kind_of?String
        # Raise something for invalid cases ?
    end
end

Well... (1) I don't want to use any global var ($labels), (2) Seperating
'records' and 'whole database' (DB) would be fine.

> Here's a little example, data file first then program file.  It first
> grabs the first line from the file, and gets the field names from it.
> Then it reads the other lines; for each one, it creates a hash (with
> the field names as keys and the values from the line as values), and
> pushes that hash onto the TSDB object itself.
> 
> # tsd.txt (tab-separated):
> field1	field2	field3
> one	two	three
> apple	orange	banana
> 
> # tsd.rb:
> class TSDB < Array
> 
>   def initialize(file)
>     concat ["Dummy 0'th element"]  # so that you can index from 1
>     fh = File.open(file, "r")
>     fields = fh.readline.chomp.split(/\t/)
> 
>     fh.readlines.each do |line|
>       data = line.chomp.split(/\t/)
>       record = {}
>       fields.each_with_index {|f,i| record[f] = data[i]}
>       push(record)
>     end
>   end
> end
> 
> t = TSDB.new("tsd.txt")
> p t[1]["field1"]          # =>  "one"
> p t[2]["field2"]          # =>  "orange"
> 
> __END__
> 
> 
> Since t[1] is a hash, you can get at its values with strings (i.e., its
> keys are strings).
> 
Nice piece of code :) Short and readable... Thanks again.

-- 
roktas