distinct texticle.rb

class Auction
  # is categorized once or twice
  # with a primary_category_id and a secondary_category_id
  # there is a Category class that acts_as_tree

  # this scope will select duplicates if auction
  # is categorized with a "brother" category because
  # it has two categorizations included in the query
  named_scope :in_category, lambda { |category_id|
    { :joins => 
      sanitize_sql_array(["JOIN categories as subcategory
        ON primary_category_id = subcategory.id
        OR secondary_category_id = subcategory.id
      JOIN categories as category
        ON subcategory.parent_id = category.id
        AND category.id = ?", category_id]) 
    }
  }

  # texticle 
  index do
    title 
    description
  end
end


# to pull results I chain the search parameters.
# first I grab the auctions in the given category:
@chain = Auction.in_category(category_id)

# calling search will use overwrite :select option
# :select => "#{table_name}.*, ts_rank_cd((#{full_text_indexes.first.to_s}), to_tsquery(#{connection.quote(term)})) as rank"
#  only search if a search value is not empty
@chain = @chain.search(text) if !text.to_s.empty?

@chain.all


# so... if I pass in :select=>"Distinct auctions.*" it will be 
# overwritten or i could reverse the order and the distinct will 
# be overwritten.  The other complication is that the 
# category_id is also an optional parameter.

# I'd love it if I could do something like this:
:select => options[:select] ? "DISTINCT " + options[:select] : "DISTINCT auctions.*"

# options is not defined - if i could check for a :select defined already... 
# how can i access the hash with which i am merging? probably not going to happen.