将 sql VS 活动记录请求与 Benchmark-ips 进行比较答案

【问题标题】：Compare sql VS active record requests with Benchmark-ips将 sql VS 活动记录请求与 Benchmark-ips 进行比较
【发布时间】：2015-08-14 07:49:09
【问题描述】：

我正在尝试将活动记录请求与我的 rails 项目中的 sql 请求进行比较。我正在使用Benchmark-ips 并撬动。

这里发生了什么：

[4] pry(main)> Benchmark.ips do |x|
[4] pry(main)*   x.report("sql request") { User.where("EXISTS (SELECT 1 FROM groups_users WHERE groups_users.user_id = users.id AND groups_users.group_id IN (?))", [8939, 8950]).where("NOT EXISTS (SELECT 1 FROM groups_users WHERE groups_users.user_id = users.id AND groups_users.group_id IN (?))", [8942]).ids }
[4] pry(main)*   x.report("active record") { (User.joins(:groups).where(groups: {id: ["8939","8950"]}) - User.joins(:groups).where(groups: {id: 8942})).map(&:id) }
[4] pry(main)*   x.compare!
[4] pry(main)* end
Calculating -------------------------------------
         sql requestNameError: uninitialized constant User
from (pry):15:in `block (2 levels) in __pry__'

我用经典的基准测试工具做了同样的事情。

[65] pry(main)> Benchmark.bm do |x|
[65] pry(main)*   x.report("sql request") { User.where("EXISTS (SELECT 1 FROM groups_users WHERE groups_users.user_id = users.id AND groups_users.group_id IN (?))", [8939, 8950]).where("NOT EXISTS (SELECT 1 FROM groups_users WHERE groups_users.user_id = users.id AND groups_users.group_id IN (?))", [8942]).ids }
[65] pry(main)*   x.report("active record") { (User.joins(:groups).where(groups: {id: ["8939","8950"]}) - User.joins(:groups).where(groups: {id: 8942})).map(&:id) }
[65] pry(main)* end
       user     system      total        real
sql request   (1.5ms)  SELECT "users"."id" FROM "users" WHERE (EXISTS (SELECT 1 FROM groups_users WHERE groups_users.user_id = users.id AND groups_users.group_id IN (8939,8950))) AND (NOT EXISTS (SELECT 1 FROM groups_users WHERE groups_users.user_id = users.id AND groups_users.group_id IN (8942)))
  0.010000   0.010000   0.020000 (  0.027799)
active record  User Load (9.4ms)  SELECT "users".* FROM "users" INNER JOIN "groups_users" ON "groups_users"."user_id" = "users"."id" INNER JOIN "groups" ON "groups"."id" = "groups_users"."group_id" WHERE "groups"."id" IN (8939, 8950)
  User Load (0.8ms)  SELECT "users".* FROM "users" INNER JOIN "groups_users" ON "groups_users"."user_id" = "users"."id" INNER JOIN "groups" ON "groups"."id" = "groups_users"."group_id" WHERE "groups"."id" = $1  [["id", 8942]]
  0.050000   0.000000   0.050000 (  0.074597)
=> [#<Benchmark::Tms:0x007ff477ce7c28 @cstime=0.0, @cutime=0.0, @label="sql request", @real=0.02779875499982154, @stime=0.010000000000000009, @total=0.02000000000000024, @utime=0.010000000000000231>,
 #<Benchmark::Tms:0x007ff473f3f7e0 @cstime=0.0, @cutime=0.0, @label="active record", @real=0.07459704099892406, @stime=0.0, @total=0.04999999999999982, @utime=0.04999999999999982>]

【问题讨论】：

另外，使用#pluck 代替#map(id) 并将其添加到第一组：.where(groups: {id: ["8939","8950"]}).pluck(:id)。另外，将您的代码包装在ActiveRecord::Base.uncached do ... end。
非常感谢 EugZol。为什么要取消缓存？费用吗？我已经重新运行了没有ids 或pluck 的请求，只是count。 active record: 127.8 i/s - 2.45x slower

标签： ruby-on-rails rails-activerecord pry

【解决方案1】：

来吧贝尼！

如果你没有User，那是因为你开始pry而不是pry-rails。

答案是简单地启动您的rails console。

然后繁荣，它的工作原理。

Comparison:
         sql request:      299.4 i/s
       active record:      123.6 i/s - 2.42x slower

=> #<Benchmark::IPS::Report:0x007fe0517c1008
 @data=nil,
 @entries=
  [#<Benchmark::IPS::Report::Entry:0x007fe04f7a9158 @ips=299.353570090315, @ips_sd=41, @iterations=1479, @label="sql request", @measurement_cycle=29, @microseconds=5037948.6083984375, @show_total_time=false>,
   #<Benchmark::IPS::Report::Entry:0x007fe050e004b0 @ips=123.57053401687592, @ips_sd=13, @iterations=612, @label="active record", @measurement_cycle=12, @microseconds=5004271.2688446045, @show_total_time=false>]>

【讨论】：