[newbie] linked hash

ruby-talk
1

i’m transalating a BeanShell prog using :
Map fieldsHash = new LinkedHashMap();
fieldsHash.put(“btl_desc”, “Description”);
fieldsHash.put(“btl_vintage”, “Millesime”);
fieldsHash.put(“btl_buy_date”, “Date achat”);
fieldsHash.put(“wcol_id”, “Coul.”);
fieldsHash.put(“btl_init_nbr”, “Qte A”);
fieldsHash.put(“btl_current_nbr”, “Qte R”);

i’ve seen the PseudoHash of Gotoken :

class PseudoHash < Hash

VERSION = “2000-12-25”

def order
@order ||= []
end

def order=(ary)
order.replace ary
end

def with_order(ary)
self.order = ary
self
end

def order_of(key)
order.index(key)
end

alias each each
alias aset []=

def []=(k, *rest)
val, set = rest.reverse
order.push k if set and not order.include? k
aset(k,val)
end

def each
rest = keys - order
order.each {|k| yield(k, self[k]) if include?(k)}
rest.each {|k| yield(k, self[k]) }
end

alias each_pair each
end

i want to emulate that LinkedHashMap into ruby without calling the java
class (it’s possible because i’m using JRuby) how could i make
modifications, if any, Gotoken’s code for that ?

···

–
yt

2

“Yvon Thoraval” yvon.thoravalNO-SPAM@free.fr schrieb im Newsbeitrag
news:1gbat4z.108b9591bhia0iN%yvon.thoravalNO-SPAM@free.fr…

i’m transalating a BeanShell prog using :
Map fieldsHash = new LinkedHashMap();
fieldsHash.put(“btl_desc”, “Description”);
fieldsHash.put(“btl_vintage”, “Millesime”);
fieldsHash.put(“btl_buy_date”, “Date achat”);
fieldsHash.put(“wcol_id”, “Coul.”);
fieldsHash.put(“btl_init_nbr”, “Qte A”);
fieldsHash.put(“btl_current_nbr”, “Qte R”);

i’ve seen the PseudoHash of Gotoken :

i want to emulate that LinkedHashMap into ruby without calling the java
class (it’s possible because i’m using JRuby) how could i make
modifications, if any, Gotoken’s code for that ?

Dunno that class but what is it that you want to achieve? Do you need a
sorted Hash? Then you can easily do

h = {}

modify h

h.sort.each do |key, val|

…

end

or even

h.sort {|a,b| a <=> b}.each do |key, val|

…

end

Kind regards

robert
3

no ordered only as it is entered, in the mean time i’ve found a solution
:

OrderedHash.rb

class OrderedHash < Hash

VERSION = “2004-03-28”

    def []=( key, value )
            ( @ordered_keys ||= [] ) << key unless has_key?( key )
            super
    end

    def each
            @ordered_keys.each do | key |
            yield key, fetch( key )
            end if @ordered_keys
    end

    def keys
            a = Array.new
            @ordered_keys.each do | key |
            a << key
            end if @ordered_keys
            return a
    end

    def join_keys(s)
            return self.keys.join(s)
    end

    def values
            a = Array.new
            @ordered_keys.each do | key |
            a << fetch(key)
            end if @ordered_keys
            return a
    end

    def join_values(s)
            return self.values.join(s)
    end

end

···

Robert Klemme bob.news@gmx.net wrote:

Dunno that class but what is it that you want to achieve? Do you need a
sorted Hash?

–
yt

4

“Yvon Thoraval” yvon.thoravalNO-SPAM@free.fr schrieb im Newsbeitrag
news:1gbf4b9.12juctd1qzju96N%yvon.thoravalNO-SPAM@free.fr…

Dunno that class but what is it that you want to achieve? Do you need
a
sorted Hash?

no ordered only as it is entered, in the mean time i’ve found a solution

Ah, I see. Alternatively you can use an LRU hash map like the one
attached. You can easily adapt this to only reflect insertion order and
not general access order.

I guess there are other implementations in the RAA.

Kind regards

robert

LRUCache.rb (1.8 KB)

···

Robert Klemme bob.news@gmx.net wrote:

5

Making a subclass of a standard class like Hash can be dangerous,
you should verify that it works for all methods, otherwise you may get
surprises:

h = OrderedHash.new
h[‘key1’] = 1
h[‘key2’] = 2
h.delete(‘key1’)
Now ‘key1’ is still in the ‘@ordered_keys’ array, and methods like
‘keys’ and ‘each’ will give you wrong results.

Regards

Karsten Meier

···

no ordered only as it is entered, in the mean time i’ve found a solution
:

OrderedHash.rb

class OrderedHash < Hash

VERSION = “2004-03-28”

    def []=( key, value )
            ( @ordered_keys ||= [] ) << key unless has_key?( key )
            super
    end

    def each
            @ordered_keys.each do | key |
            yield key, fetch( key )
            end if @ordered_keys
    end

    def keys
            a = Array.new
            @ordered_keys.each do | key |
            a << key
            end if @ordered_keys
            return a
    end

    def join_keys(s)
            return self.keys.join(s)
    end

    def values
            a = Array.new
            @ordered_keys.each do | key |
            a << fetch(key)
            end if @ordered_keys
            return a
    end

    def join_values(s)
            return self.values.join(s)
    end

end

6

ok, tanxs for your help

Yvon

···

Robert Klemme bob.news@gmx.net wrote:

Ah, I see. Alternatively you can use an LRU hash map like the one
attached. You can easily adapt this to only reflect insertion order and
not general access order.

I guess there are other implementations in the RAA.

–
yt

7

fine °;=) tanxs… i don’t use “delete” today but…

···

Karsten Meier discussionruby@internetlabor.de wrote:

h = OrderedHash.new
h[‘key1’] = 1
h[‘key2’] = 2
h.delete(‘key1’)
Now ‘key1’ is still in the ‘@ordered_keys’ array, and methods like
‘keys’ and ‘each’ will give you wrong results.

–
yt