processing large csv files in ruby

gistfile1.rb

class ZendeskTicketsJob
  extend Resque::Plugins::ExponentialBackoff
  @queue = :low
  FIELDS = ['zendesk_id', 'requester_id', 'assignee_id', 'group', 'subject', 'tags', 'status', 'priority', 'via', 'ticket_type', 'created_at', 'assigned_at', 'solved_at', 'resolution_time', 'satisfaction', 'group_stations', 'assignee_stations', 'reopens', 'replies', 'first_reply_time_in_minutes', 'first_reply_time_in_minutes_within_business_hours', 'first_resolution_time_in_minutes', 'first_resolution_time_in_minutes_within_business_hours', 'full_resolution_time_in_minutes', 'full_resolution_time_in_minutes_within_business_hours', 'agent_wait_time_in_minutes', 'agent_wait_time_in_minutes_within_business_hours', 'requester_wait_time_in_minutes', 'requester_wait_time_in_minutes_within_business_hours', 'reservation_code', 'requires_manual_closing']

  def self.perform(url)
    `rm /tmp/zendesk_tickets*`
    `wget #{url} -O /tmp/zendesk_tickets.csv.zip`
    `unzip -p /tmp/zendesk_tickets.csv.zip > /tmp/zendesk_tickets.csv`
    
    # IO.foreach doesn't read the entire file into memory at once, which is good since a standard FasterCSV.parse on this file can take an hour or more
    lines = []
    IO.foreach('/tmp/zendesk_tickets.csv') do |line|
      lines << line
      if lines.size >= 1000
        lines = FasterCSV.parse(lines.join) rescue next
        store lines
        lines = []
      end
    end
    store lines
  end
  
  def self.store lines
    return if lines.blank?
    puts "Storing #{lines.size} rows..."
    connection.execute("
      INSERT INTO zendesk_tickets (#{fields_to_sql FIELDS})
      VALUES #{data_to_sql lines}
      ON DUPLICATE KEY
      UPDATE #{fields_to_update_sql FIELDS[1..-1]}
    ")
  end

  def self.fields_to_sql fields
    fields.collect{|f| connection.quote_column_name(f) }.join(',')
  end

  def self.data_to_sql lines
    data = lines.collect do |v|
      # ... some manipulation of the data into our own formats
      v
    end.compact
    
    data.collect do |row|
      "(#{row.collect{|f| connection.quote(f)}.join(',')})"
    end.join(', ')
  end
  
  def self.connection
    @connection ||= ZendeskTicket.new.connection
  end

  def self.fields_to_update_sql fields
    fields.collect do |f|
      qcn = connection.quote_column_name(f)
      "#{qcn} = VALUES(#{qcn})"
    end.join(', ')
  end

end

What kind of performance did this yield for you?