How to (Massively) Reduce The Number of SQL Queries in Rails App

How to (massively) reduce the number of SQL queries in Rails app?

Whenever revenue_between is called, it fetches the payments in the given time range and the associated invoices and items from the db. Since the time ranges have lot of overlap (month, quarter, year), same records are being fetched over and over again.

I think it is better to fetch all the payments of the user once, then filter and summarize them in Ruby.

To implement, change the revenue_between method as follows:

def revenue_between(range, kind) 
   #store the all the payments as instance variable to avoid duplicate queries
   @payments_with_invoice ||= payments.includes(:invoice => :items).all
   @payments_with_invoice.select{|x| range.cover? x.created_at}.sum(&:"#{kind}_amount") 
end

This would eager load all the payments along with associated invoices and items.

Also change the invoice summation methods so that it uses the eager loaded items

class Invoice < ActiveRecord::Base

   def total
     items.map(&:total).sum
   end

   def subtotal
     items.map(&:subtotal).sum
   end

   def total_tax
     items.map(&:total_tax).sum
   end

end

How can I prevent my Rails query from maxing out memory

The issue was with sending tasks to Resque. Thanks to @user2864740 for having this idea.

First I tested by using an empty find_each block, and that didn't cause memory growth. Then I tried a find_each block that created an ActiveRecord object, and that didn't cause memory growth.

Finally I tested doing a find_each over records and sending off a task to resque. This alone caused the memory to steadily climb and not fall. I should have also mentioned that I'm using the ResqueDelayable gem to dispatch tasks to resque, and it might be implicated by this.

auto-filtering all activerecord queries in rails app

use default_scope for this

http://apidock.com/rails/ActiveRecord/Base/default_scope/class

in your case s.th. like

default_scope where("product_id in (?)",[allowed_products_array])

you can extract it into module
and include into every model you'd like to have this functionality.

Implementing a filter for a SQL query in a Rails app

I would consider creating a materialized view of the union of the two tables.

CREATE MATERIALIZED VIEW unified_notifications AS
   SELECT 
      sn.id, 
      sellers.name, 
      sn.seller_id AS member_id, 
      sn.title, 
      sn.body, 
      sn.created_at, 
      sn.is_new, 
      sellers.platform_id, 
      sn.pending_action_id, 
      'seller_notifications' AS type 
   FROM seller_notifications sn
   INNER JOIN sellers ON sellers.id = sn.seller_id
   UNION ALL
   SELECT 
     un.id, 
     users.name, 
     un.user_id AS member_id, 
     un.title, 
     un.body, 
     un.created_at, 
     un.is_new, 
     users.platform_id, 
     un.pending_action_id, 
     'user_notifications' AS type 
   FROM user_notifications un
   INNER JOIN users ON users.id = un.user_id

The scenic gem can be used to create views with Rails migrations. Note that you need to switch to SQL schema dumps.

Since the view behaves like a table (at least from the perspective of ActiveRecord) you can just create an ActiveRecord model and query it like it was a single table:

class UnifiedNotification < ApplicationRecord
  def readonly?
    true
   end
end

UnifiedNotification.where(
  platform_id: @user.platform_id,
  # ...
)

Rails3: SQL execution with hash substitution like .where()

query = ActiveRecord::Base.connection.raw_connection.prepare("INSERT INTO users (name) VALUES(:name)")
query.execute(:name => 'test_name')
query.close

Counting the number of queries performed

I think you answered your own question by mentioning assert_queries, but here goes:

I would recommend taking a look at the code behind assert_queries and using that to build your own method which you can use to count queries. The main magic involved here is this line:

ActiveSupport::Notifications.subscribe('sql.active_record', SQLCounter.new)

I had a bit of a tinker this morning and ripped out the parts of ActiveRecord that do the query counting and came up with this:

module ActiveRecord
  class QueryCounter
    cattr_accessor :query_count do
      0
    end

    IGNORED_SQL = [/^PRAGMA (?!(table_info))/, /^SELECT currval/, /^SELECT CAST/, /^SELECT @@IDENTITY/, /^SELECT @@ROWCOUNT/, /^SAVEPOINT/, /^ROLLBACK TO SAVEPOINT/, /^RELEASE SAVEPOINT/, /^SHOW max_identifier_length/]

    def call(name, start, finish, message_id, values)
      # FIXME: this seems bad. we should probably have a better way to indicate
      # the query was cached
      unless 'CACHE' == values[:name]
        self.class.query_count += 1 unless IGNORED_SQL.any? { |r| values[:sql] =~ r }
      end
    end
  end
end

ActiveSupport::Notifications.subscribe('sql.active_record', ActiveRecord::QueryCounter.new)

module ActiveRecord
  class Base
    def self.count_queries(&block)
      ActiveRecord::QueryCounter.query_count = 0
      yield
      ActiveRecord::QueryCounter.query_count
    end
  end
end

You will be able to reference the ActiveRecord::Base.count_queries method anywhere. Pass it a block wherein your queries are run and it will return the number of queries that have been executed:

ActiveRecord::Base.count_queries do
  Ticket.first
end

Returns "1" for me. To make this work: put it in a file at lib/active_record/query_counter.rb and require it in your config/application.rb file like this:

require 'active_record/query_counter'

Hey presto!

---

A little bit of explanation probably is required. When we call this line:

    ActiveSupport::Notifications.subscribe('sql.active_record', ActiveRecord::QueryCounter.new)

We hook into Rails 3's little notifications framework. It's a shiny little addition to the latest major version of Rails that nobody really knows about. It allows us to subscribe to notifications of events within Rails by using the subscribe method. We pass in the event we want to subscribe to as the first argument then any object that responds to call as the second.

In this case when a query is executed our little query counter will dutifully increment the ActiveRecord::QueryCounter.query_count variable, but only for the real queries.

Anyway, this was fun. I hope it comes useful to you.

ActiveRecord Rails - create a batch of empty records

Take a look at the activerecord-import gem. It handles bulk inserts like you described, while keeping the number of ActiveRecord queries to a minimum. And it's probably a good idea to do this in a background job since you are creating so many records.

---

Related Topics