如何对mongodb/mongoid脚本进行基准测试，以比较两种不同的查询技术 [英] How to benchmarking mongodb/mongoid scripts, for comparing two different query techniques

问题描述

您是否对如何在两种不同的monongoid/mongodb查询实现中测试性能提出建议?

Do you have a suggestion on how to test performances over two different mongoid/mongodb query implementations?

要比较的实现与以前的有关问与答" ，为简便起见，我将再次在此报告:

The implementation to compare, are related to a previous "Q & A" but I'll report those here again, for brevity :

第一个代码:

Competitor  = Struct.new(:html_url, :description, :user)
competitors = []

User.all.map do |user|
  user.watchlists.all.map do |wl|
    if wl.tags_array == ["ruby", "web", "framework"]
        competitors << Competitor.new(wl.html_url, wl.description, wl.user.nickname)
    end
  end
end

与第二个代码:

Competitor = Struct.new(:html_url, :description, :user)
competitors = []

User.where('watchlists.tags_array' => %w[ruby web framework]).only(:nickname, :watchlists).each do |u|
  u.watchlists.where(:tags_array => %w[ruby web framework]).each do |wl|
    competitors << Competitor.new(wl.html_url, wl.description, u.nickname)
  end
end

为完善起见，

下划线数据模型为:

class User
  include Mongoid::Document

  field :nickname
  embeds_many :watchlists
end

class Watchlist
 include Mongoid::Document

 field :html_url
 field :description
 field :tags_array, type: Array
 embedded_in :user
end

目前进行基准测试&的方式是什么?比较这两个代码的性能?

What's the current way for benchmarking & comparing performances on that two codes ?

推荐答案

直到现在，我还是这样找到的，但是如果您知道更好的方法，请也回答...

Until now, I found this way but if you know something better please answer too ...

第一个代码:

1.9.2p290 :038 >   Competitor  = Struct.new(:html_url, :description, :user)
1.9.2p290 :039 >   time = Benchmark.measure do
1.9.2p290 :040 >     competitors = []
1.9.2p290 :041?>   
1.9.2p290 :042 >     User.all.map do |user|
1.9.2p290 :043 >         user.watchlists.all.map do |wl|
1.9.2p290 :044 >             if wl.tags_array == ["ruby", "web", "framework"]
1.9.2p290 :045?>                 competitors << Competitor.new(wl.html_url, wl.description, wl.user.nickname)
1.9.2p290 :046?>             end
1.9.2p290 :047?>         end
1.9.2p290 :048?>     end
1.9.2p290 :049?>   end
MONGODB (34ms) heroku_app1707530['users'].find({})
MONGODB [DEBUG] cursor.refresh() for cursor 2492208785546818168
MONGODB [DEBUG] cursor.refresh() for cursor 2492208785546818168
 =>   1.390000   0.010000   1.400000 (  1.400842)

1.9.2p290 :050 > 

与第二个代码:

1.9.2p290 :049 > Competitor  = Struct.new(:html_url, :description, :user)
1.9.2p290 :050 > time = Benchmark.measure do
1.9.2p290 :051 >     competitors = []
1.9.2p290 :052?>   
1.9.2p290 :053 >     User.where('watchlists.tags_array' => %w[ruby web framework]).only(:nickname, :watchlists).each do |u|
1.9.2p290 :054 >         u.watchlists.where(:tags_array => %w[ruby web framework]).each do |wl|
1.9.2p290 :055 >             competitors << Competitor.new(wl.html_url, wl.description, u.nickname)
1.9.2p290 :056?>         end
1.9.2p290 :057?>     end
1.9.2p290 :058?>   end
MONGODB (185ms) heroku_app1707530['users'].find({"watchlists.tags_array"=>["ruby", "web", "framework"]}, {:_type=>1, :nickname=>1, :watchlists=>1})
MONGODB [DEBUG] cursor.refresh() for cursor 2195378613558492020
 =>   0.440000   0.000000   0.440000 (  0.456714)

1.9.2p290 :059 > 

但更好:

第一个代码:

1.9.2p290 :157 > Competitor  = Struct.new(:html_url, :description, :user)
1.9.2p290 :158 > competitors = []
 => [] 
1.9.2p290 :159 > Benchmark.bm(10) do |x|
1.9.2p290 :160 >     x.report("first:") do
1.9.2p290 :161 >       User.all.map do |user|
1.9.2p290 :162 >           user.watchlists.all.map do |wl|
1.9.2p290 :163 >               if wl.tags_array == ["ruby", "web", "framework"]
1.9.2p290 :164?>                   competitors << Competitor.new(wl.html_url, wl.description, wl.user.nickname)
1.9.2p290 :165?>               end
1.9.2p290 :166?>           end
1.9.2p290 :167?>       end
1.9.2p290 :168?>     end
1.9.2p290 :169?>   x.report("second:") do
1.9.2p290 :170 >       User.all.map do |user|
1.9.2p290 :171 >           user.watchlists.all.map do |wl|
1.9.2p290 :172 >               if wl.tags_array == ["ruby", "web", "framework"]
1.9.2p290 :173?>                   competitors << Competitor.new(wl.html_url, wl.description, wl.user.nickname)
1.9.2p290 :174?>               end
1.9.2p290 :175?>           end
1.9.2p290 :176?>       end
1.9.2p290 :177?>     end
1.9.2p290 :178?>   x.report("third:") do
1.9.2p290 :179 >       User.all.map do |user|
1.9.2p290 :180 >           user.watchlists.all.map do |wl|
1.9.2p290 :181 >               if wl.tags_array == ["ruby", "web", "framework"]
1.9.2p290 :182?>                   competitors << Competitor.new(wl.html_url, wl.description, wl.user.nickname)
1.9.2p290 :183?>               end
1.9.2p290 :184?>           end
1.9.2p290 :185?>       end
1.9.2p290 :186?>     end
1.9.2p290 :187?>   end
                user     system      total        real
first:    MONGODB (30ms) heroku_app1707530['users'].find({})
MONGODB [DEBUG] cursor.refresh() for cursor 6320857008182747446
MONGODB [DEBUG] cursor.refresh() for cursor 6320857008182747446
  1.460000   0.010000   1.470000 (  1.475545)
second:   MONGODB (24ms) heroku_app1707530['users'].find({})
MONGODB [DEBUG] cursor.refresh() for cursor 8580701579081246457
MONGODB [DEBUG] cursor.refresh() for cursor 8580701579081246457
  1.470000   0.010000   1.480000 (  1.494812)
third:    MONGODB (24ms) heroku_app1707530['users'].find({})
MONGODB [DEBUG] cursor.refresh() for cursor 6472818135140756688
MONGODB [DEBUG] cursor.refresh() for cursor 6472818135140756688
  1.490000   0.010000   1.500000 (  1.505000)
 => true 
1.9.2p290 :188 > 

与第二个代码:

1.9.2p290 :065 > Competitor  = Struct.new(:html_url, :description, :user)
1.9.2p290 :066 > competitors = []
 => [] 
1.9.2p290 :067 > Benchmark.bm(10) do |x|
1.9.2p290 :068 >     x.report("first:") {User.where('watchlists.tags_array' => %w[ruby web framework]).only(:nickname, :watchlists).each{|u|u.watchlists.where(:tags_array => %w[ruby web framework]).each{|wl|competitors << Competitor.new(wl.html_url, wl.description, u.nickname)}}}
1.9.2p290 :069?>   x.report("second:") {User.where('watchlists.tags_array' => %w[ruby web framework]).only(:nickname, :watchlists).each{|u|u.watchlists.where(:tags_array => %w[ruby web framework]).each{|wl|competitors << Competitor.new(wl.html_url, wl.description, u.nickname)}}}
1.9.2p290 :070?>   x.report("third:") {User.where('watchlists.tags_array' => %w[ruby web framework]).only(:nickname, :watchlists).each{|u|u.watchlists.where(:tags_array => %w[ruby web framework]).each{|wl|competitors << Competitor.new(wl.html_url, wl.description, u.nickname)}}}
1.9.2p290 :071?>   end
                user     system      total        real
first:    MONGODB (163ms) heroku_app1707530['users'].find({"watchlists.tags_array"=>["ruby", "web", "framework"]}, {:_type=>1, :nickname=>1, :watchlists=>1})
MONGODB [DEBUG] cursor.refresh() for cursor 7239021952645827000
  0.330000   0.000000   0.330000 (  0.380199)
second:   MONGODB (164ms) heroku_app1707530['users'].find({"watchlists.tags_array"=>["ruby", "web", "framework"]}, {:_type=>1, :nickname=>1, :watchlists=>1})
MONGODB [DEBUG] cursor.refresh() for cursor 4816095738351422260
  0.320000   0.010000   0.330000 (  0.381196)
third:    MONGODB (125ms) heroku_app1707530['users'].find({"watchlists.tags_array"=>["ruby", "web", "framework"]}, {:_type=>1, :nickname=>1, :watchlists=>1})
MONGODB [DEBUG] cursor.refresh() for cursor 5014359782446173361
  0.390000   0.010000   0.400000 (  0.397241)
 => true 
1.9.2p290 :072 > 

然后，第二个代码的确速度更快，CPU消耗更少，...， 感谢 rubish 为第二个代码"

then second code results really faster and less cpu consuming ..., thanks to rubish for the "second code"

这篇关于如何对mongodb/mongoid脚本进行基准测试，以比较两种不同的查询技术的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！