Refactor methods that rely on instance variable

How can refactor these methods so that they don't rely on using an instance variable (@conference_facets)?

def count_conf_facets(events)
  @conference_facets =开发者_Python百科 {}
  for event in events do
    home_team_conf = event.home_team_conf
    away_team_conf = event.away_team_conf
    increment_conf_facet(home_team_conf)
    if home_team_conf.id != away_team_conf.id then
      increment_conf_facet(away_team_conf)
    end
  end
end

def increment_conf_facet(conference)
  if @conference_facets[conference.id].nil? then
    @conference_facets[conference.id] = 0
  end
  @conference_facets[conference.id] += 1
end

If these facets are stored in a database, you can get the count right out of the database, which is much more efficient than looping through all the events, although you may have to play around with your group_by and join criteria, as well as your data modeling.

Anyway, to get rid of having to use an instance variable, you can move all of this logic into a single function, or play around with what you're passing between functions. For example

def count_conf_facets(events)
  counts = {}
  events.each do |event|
    [event.home_team_conf.id, event.away_team_conf.id].uniq.each do |conf_id|
      counts[conf_id] ||= 0
      counts[conf_id] += 1
    end
  end

  counts
end