From: ts <decoux@...>
Date: 2001-08-16T01:02:14+09:00
Subject: [ruby-talk:19795] Re: 2 iterator questions

>>>>> "M" == Mark Slagell <ms@iastate.edu> writes:

M> 2. The conversion of a block to a proc inside an iterator(sec. 5.3) is
M> interesting... I'm curious what use it might have though.  Any thoughts?

 Well, there are many examples. In bdb I use it to implement secondary
 indexes. The block is called each time a (key, value) is inserted in the
 db file to create the second index.

pigeon% cat b.rb
#!/usr/bin/ruby
$LOAD_PATH.unshift "src"
require 'bdb'

cat = Struct.new("Cat", :name, :age, :life)

bdb = BDB::Btree.open "tmp/aa", nil, BDB::CREATE, "marshal" => true
aux = BDB::Btree.open "tmp/bb", nil, BDB::CREATE, 
   "set_flags" => BDB::DUP | BDB::DUPSORT, "marshal" => true
bdb.associate(aux) { |aux, key, value| value.life }
bdb["aa"] = cat.new "cat1", 12, 7
bdb["ab"] = cat.new "cat2", 12, 6
bdb["ac"] = cat.new "cat3", 12, 5
bdb["ad"] = cat.new "cat4", 12, 7
aux.each do |k, v|
   p "key :#{k} -- value #{v.inspect}"
end
pigeon% 

pigeon% b.rb
"key :5 -- value #<Struct::Cat name=\"cat3\", age=12, life=5>"
"key :6 -- value #<Struct::Cat name=\"cat2\", age=12, life=6>"
"key :7 -- value #<Struct::Cat name=\"cat1\", age=12, life=7>"
"key :7 -- value #<Struct::Cat name=\"cat4\", age=12, life=7>"
pigeon% 


Guy Decoux