tatyree-query-analyzer 0.1.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/README.rdoc +108 -0
- data/Rakefile +12 -0
- data/init.rb +1 -0
- data/lib/query_analyzer.rb +56 -0
- data/query-analyzer.gemspec +31 -0
- metadata +64 -0
data/README.rdoc
ADDED
@@ -0,0 +1,108 @@
|
|
1
|
+
= Query Analyzer Plugin for MySQL on Rails
|
2
|
+
|
3
|
+
A gem version of the Query Analyzer plugin. Modified by Todd Tyree.
|
4
|
+
|
5
|
+
MODIFIED by John Eberly originally take from http://svn.nfectio.us
|
6
|
+
|
7
|
+
The Query Analyzer plugin will expand the usability of your log files
|
8
|
+
by providing query analysis using the MySQL query execution plan. Each SQL
|
9
|
+
select query will be 'EXPLAIN'ed and added to the log files right below
|
10
|
+
the original query.
|
11
|
+
|
12
|
+
Using this plugin and a good understanding of the results, you will be
|
13
|
+
able to analyze and optimize the queries your application is making.
|
14
|
+
|
15
|
+
Refer to http://www.mysql.org/doc/refman/5.0/en/explain.html for more
|
16
|
+
information on understanding the results.
|
17
|
+
|
18
|
+
= Compatibility
|
19
|
+
|
20
|
+
It only works with MySQL at the moment.
|
21
|
+
|
22
|
+
= Installation
|
23
|
+
|
24
|
+
sudo gem install git://github.com/tatyree/query-analyzer.git
|
25
|
+
|
26
|
+
In your environment.rb, *after* the Rails::Initializer block:
|
27
|
+
|
28
|
+
require 'query_analyzer'
|
29
|
+
|
30
|
+
*NOTE THE UNDERSCORE!* +require 'query-analyzer'+ won't find it.
|
31
|
+
|
32
|
+
= Example Use
|
33
|
+
|
34
|
+
Here is a real life usage of the plugin that detected the omission of indexes on
|
35
|
+
a table. In this case, it was a join table and the keys didn't have indexes (silly me!).
|
36
|
+
Names have been changed to protect the innocent (and make it fit 80 columns)
|
37
|
+
|
38
|
+
# development.log
|
39
|
+
|
40
|
+
P Load (0.008669)
|
41
|
+
=> SELECT p.* FROM p INNER JOIN d ON p.id = d.p_id WHERE (d.p_id = 2 AND ((d.type = 'P')))
|
42
|
+
|
43
|
+
Analyzing P Load
|
44
|
+
|
45
|
+
select_type | key_len | type | Extra | id | possible_keys | rows | table | ref | key
|
46
|
+
----------------------------------------------------------------------------------------------------
|
47
|
+
SIMPLE | | ALL | Using where | 1 | | 74 | d | |
|
48
|
+
SIMPLE | 4 | eq_ref | Using where | 1 | PRIMARY | 1 | p | d.p_id | PRIMARY
|
49
|
+
|
50
|
+
|
51
|
+
= Analyze the results
|
52
|
+
|
53
|
+
Looking at the results of the execution plan, we can see that the lookup in
|
54
|
+
the d table is missing an index (possible_keys=null) and performed a full
|
55
|
+
table scan (type=ALL) to satisfy the WHERE condition. In this case, there was only one
|
56
|
+
row that matched the condition in the table, but MySQL still had to search all 74 rows
|
57
|
+
in the table to find it, a key indicator of a missing
|
58
|
+
or malformed index(es). Once it has pulled all the records to satisfy the WHERE, it then
|
59
|
+
starts the p table join. This time, it was able to match d.p_id to p.id using
|
60
|
+
the PRIMARY key on the p table. The type=eq_ref indicates a 1 to 1 match against a primary
|
61
|
+
or unique column.
|
62
|
+
|
63
|
+
Lets add some indexes to the join table and see if we can cut that full table scan down in the
|
64
|
+
number of rows it needs to search.
|
65
|
+
|
66
|
+
#> script/generate Migration AddIndexesToD
|
67
|
+
|
68
|
+
# file: 005_add_indexes_to_d
|
69
|
+
class AddIndexesToD < ActiveRecord::Migration
|
70
|
+
def self.up
|
71
|
+
add_index :d, [ :p_id, :type ]
|
72
|
+
add_index :d, :type
|
73
|
+
end
|
74
|
+
|
75
|
+
def self.remove
|
76
|
+
remove_index :d, [ :p_id, :type ]
|
77
|
+
remove_index :d, :type
|
78
|
+
end
|
79
|
+
end
|
80
|
+
|
81
|
+
|
82
|
+
Now that we have an index on the foreign_key column and type, lets re-run the query and
|
83
|
+
see if we got rid of that full table scan.
|
84
|
+
|
85
|
+
# development.log
|
86
|
+
|
87
|
+
P Load (0.009011)
|
88
|
+
=> SELECT p.* FROM p INNER JOIN d ON p.id = d.p_id WHERE (d.p_id = 2 AND ((d.type = 'P')))
|
89
|
+
|
90
|
+
Analyzing P Load
|
91
|
+
|
92
|
+
select_type | key_len | type | Extra | id | possible_keys | rows | table | ref | key
|
93
|
+
------------------------------------------------------------------------------------------------------------------------------
|
94
|
+
SIMPLE | 255 | ref | Using where | 1 | d_p_id_type_index,d_type_index | 1 | d | const | d_p_id_type_index
|
95
|
+
SIMPLE | 4 | eq_ref | Using where | 1 | PRIMARY | 1 | p | d.p_id | PRIMARY
|
96
|
+
|
97
|
+
Okay. Now MySQL is using an index satisfy the WHERE condition. Using the index, it was able to
|
98
|
+
find the single row that matched, preventing the full table scan.
|
99
|
+
|
100
|
+
|
101
|
+
Credits:
|
102
|
+
The extension of the Array class for printing the columnized records was originally
|
103
|
+
written by Peter Cooper who adapted it from Courtenay from #caboose.
|
104
|
+
|
105
|
+
http://www.rubyinside.com/columnized-text-datasets-in-rails-71.html
|
106
|
+
http://habtm.com/articles/2006/06/10/pretty-tables-for-ruby-objects
|
107
|
+
|
108
|
+
Released under the MIT license (download your own if you need it)
|
data/Rakefile
ADDED
@@ -0,0 +1,12 @@
|
|
1
|
+
require 'rubygems'
|
2
|
+
require 'rake'
|
3
|
+
require 'echoe'
|
4
|
+
|
5
|
+
Echoe.new('query-analyzer', '0.1.0') do |p|
|
6
|
+
p.description = "Run explain on all the sql queries rails generates."
|
7
|
+
p.url = "http://github.com/tatyree/cookieless_sessions"
|
8
|
+
p.author = "John Eberly and Todd Tyree"
|
9
|
+
p.email = "todd@snappl.co.uk"
|
10
|
+
p.ignore_pattern = ["tmp/*", "script/*"]
|
11
|
+
p.development_dependencies = []
|
12
|
+
end
|
data/init.rb
ADDED
@@ -0,0 +1 @@
|
|
1
|
+
require 'query_analyzer'
|
@@ -0,0 +1,56 @@
|
|
1
|
+
class Array
|
2
|
+
protected
|
3
|
+
def qa_columnized_row(fields, sized)
|
4
|
+
row = []
|
5
|
+
fields.each_with_index do |f, i|
|
6
|
+
row << sprintf("%0-#{sized[i]}s", f.to_s)
|
7
|
+
end
|
8
|
+
row.join(' | ')
|
9
|
+
end
|
10
|
+
|
11
|
+
public
|
12
|
+
|
13
|
+
def qa_columnized
|
14
|
+
sized = {}
|
15
|
+
self.each do |row|
|
16
|
+
row.values.each_with_index do |value, i|
|
17
|
+
sized[i] = [sized[i].to_i, row.keys[i].length, value.to_s.length].max
|
18
|
+
end
|
19
|
+
end
|
20
|
+
|
21
|
+
table = []
|
22
|
+
table << qa_columnized_row(self.first.keys, sized)
|
23
|
+
table << '-' * table.first.length
|
24
|
+
self.each { |row| table << qa_columnized_row(row.values, sized) }
|
25
|
+
table.join("\n ") # Spaces added to work with format_log_entry
|
26
|
+
end
|
27
|
+
end
|
28
|
+
|
29
|
+
|
30
|
+
# TODO: Postgres? Oracle? Firebird?
|
31
|
+
module ActiveRecord
|
32
|
+
module ConnectionAdapters
|
33
|
+
class MysqlAdapter < AbstractAdapter
|
34
|
+
private
|
35
|
+
alias_method :select_without_analyzer, :select
|
36
|
+
|
37
|
+
def select(sql, name = nil)
|
38
|
+
query_results = select_without_analyzer(sql, name)
|
39
|
+
|
40
|
+
# I went back to showing all queries. This makes for a heavy log file, but you pick up
|
41
|
+
# things like filesorts (which the latest plugin wouldn't have caught).
|
42
|
+
|
43
|
+
if @logger and @logger.level == Logger::DEBUG
|
44
|
+
@logger.debug(
|
45
|
+
ActiveRecord::Base.silence do
|
46
|
+
format_log_entry("\033[1;34mAnalyzing #{name} \033[0m\n\n",
|
47
|
+
"\033[1;34m#{select_without_analyzer("explain #{sql}", name).qa_columnized} \033[0m\n\n"
|
48
|
+
)
|
49
|
+
end
|
50
|
+
) if sql =~ /^select/i
|
51
|
+
end
|
52
|
+
query_results
|
53
|
+
end
|
54
|
+
end
|
55
|
+
end
|
56
|
+
end
|
@@ -0,0 +1,31 @@
|
|
1
|
+
# -*- encoding: utf-8 -*-
|
2
|
+
|
3
|
+
Gem::Specification.new do |s|
|
4
|
+
s.name = %q{query-analyzer}
|
5
|
+
s.version = "0.1.0"
|
6
|
+
|
7
|
+
s.required_rubygems_version = Gem::Requirement.new(">= 1.2") if s.respond_to? :required_rubygems_version=
|
8
|
+
s.authors = ["John Eberly and Todd Tyree"]
|
9
|
+
s.date = %q{2009-02-07}
|
10
|
+
s.description = %q{Run explain on all the sql queries rails generates.}
|
11
|
+
s.email = %q{todd@snappl.co.uk}
|
12
|
+
s.extra_rdoc_files = ["lib/query_analyzer.rb", "README.rdoc"]
|
13
|
+
s.files = ["init.rb", "lib/query_analyzer.rb", "Rakefile", "README.rdoc", "Manifest", "query-analyzer.gemspec"]
|
14
|
+
s.has_rdoc = true
|
15
|
+
s.homepage = %q{http://github.com/tatyree/cookieless_sessions}
|
16
|
+
s.rdoc_options = ["--line-numbers", "--inline-source", "--title", "Query-analyzer", "--main", "README.rdoc"]
|
17
|
+
s.require_paths = ["lib"]
|
18
|
+
s.rubyforge_project = %q{query-analyzer}
|
19
|
+
s.rubygems_version = %q{1.3.1}
|
20
|
+
s.summary = %q{Run explain on all the sql queries rails generates.}
|
21
|
+
|
22
|
+
if s.respond_to? :specification_version then
|
23
|
+
current_version = Gem::Specification::CURRENT_SPECIFICATION_VERSION
|
24
|
+
s.specification_version = 2
|
25
|
+
|
26
|
+
if Gem::Version.new(Gem::RubyGemsVersion) >= Gem::Version.new('1.2.0') then
|
27
|
+
else
|
28
|
+
end
|
29
|
+
else
|
30
|
+
end
|
31
|
+
end
|
metadata
ADDED
@@ -0,0 +1,64 @@
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
2
|
+
name: tatyree-query-analyzer
|
3
|
+
version: !ruby/object:Gem::Version
|
4
|
+
version: 0.1.0
|
5
|
+
platform: ruby
|
6
|
+
authors:
|
7
|
+
- John Eberly and Todd Tyree
|
8
|
+
autorequire:
|
9
|
+
bindir: bin
|
10
|
+
cert_chain: []
|
11
|
+
|
12
|
+
date: 2009-02-07 00:00:00 -08:00
|
13
|
+
default_executable:
|
14
|
+
dependencies: []
|
15
|
+
|
16
|
+
description: Run explain on all the sql queries rails generates.
|
17
|
+
email: todd@snappl.co.uk
|
18
|
+
executables: []
|
19
|
+
|
20
|
+
extensions: []
|
21
|
+
|
22
|
+
extra_rdoc_files:
|
23
|
+
- lib/query_analyzer.rb
|
24
|
+
- README.rdoc
|
25
|
+
files:
|
26
|
+
- init.rb
|
27
|
+
- lib/query_analyzer.rb
|
28
|
+
- Rakefile
|
29
|
+
- README.rdoc
|
30
|
+
- Manifest
|
31
|
+
- query-analyzer.gemspec
|
32
|
+
has_rdoc: true
|
33
|
+
homepage: http://github.com/tatyree/cookieless_sessions
|
34
|
+
post_install_message:
|
35
|
+
rdoc_options:
|
36
|
+
- --line-numbers
|
37
|
+
- --inline-source
|
38
|
+
- --title
|
39
|
+
- Query-analyzer
|
40
|
+
- --main
|
41
|
+
- README.rdoc
|
42
|
+
require_paths:
|
43
|
+
- lib
|
44
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
45
|
+
requirements:
|
46
|
+
- - ">="
|
47
|
+
- !ruby/object:Gem::Version
|
48
|
+
version: "0"
|
49
|
+
version:
|
50
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
51
|
+
requirements:
|
52
|
+
- - ">="
|
53
|
+
- !ruby/object:Gem::Version
|
54
|
+
version: "1.2"
|
55
|
+
version:
|
56
|
+
requirements: []
|
57
|
+
|
58
|
+
rubyforge_project: query-analyzer
|
59
|
+
rubygems_version: 1.2.0
|
60
|
+
signing_key:
|
61
|
+
specification_version: 2
|
62
|
+
summary: Run explain on all the sql queries rails generates.
|
63
|
+
test_files: []
|
64
|
+
|