Skip to content

Commit 3356bcd

Browse files
rodberiamcal
rodber
authored and
iamcal
committed
add code blocks
1 parent 9b56b6b commit 3356bcd

File tree

1 file changed

+76
-71
lines changed

1 file changed

+76
-71
lines changed

‎README.md‎

Lines changed: 76 additions & 71 deletions
Original file line numberDiff line numberDiff line change
@@ -7,102 +7,110 @@ This library takes MySQL `CREATE TABLE` statements and returns a data structure
77
MySQL syntax [version 5.7](https://dev.mysql.com/doc/refman/5.7/en/create-table.html) is supported.
88
This library does not try to validate input - the goal is to deconstruct valid `CREATE TABLE` statements.
99

10-
1110
## Installation
1211

1312
You can install this package using composer. To add it to your `composer.json`:
1413

15-
composer require iamcal/sql-parser
14+
```plain
15+
composer require iamcal/sql-parser
16+
```
1617

1718
You can then load it using the composer autoloader:
1819

19-
require_once 'vendor/autoload.php';
20-
use iamcal\SQLParser;
20+
```php
21+
require_once 'vendor/autoload.php';
22+
use iamcal\SQLParser;
2123

22-
$parser = new SQLParser();
24+
$parser = new SQLParser();
25+
```
2326

2427
If you don't use composer, you can skip the autoloader and include `src/SQLParser.php` directly.
2528

26-
2729
## Usage
2830

2931
To extract the tables defined in SQL:
3032

31-
$parser = new SQLParser();
32-
$parser->parse($sql);
33+
```php
34+
$parser = new SQLParser();
35+
$parser->parse($sql);
3336

34-
print_r($parser->tables);
37+
print_r($parser->tables);
38+
```
3539

36-
The `tables` property is an array of tables, each of which is a nested array structure defining the
40+
The `tables` property is an array of tables, each of which is a nested array structure defining the
3741
table's structure:
3842

39-
CREATE TABLE `achievements_counts` (
40-
`achievement_id` int(10) unsigned NOT NULL,
41-
`num_players` int(10) unsigned NOT NULL,
42-
`date_updated` int(10) unsigned NOT NULL,
43-
PRIMARY KEY (`achievement_id`)
44-
) ENGINE=InnoDB DEFAULT CHARSET=utf8;
45-
46-
47-
[
48-
'achievements_counts' => [
49-
'name' => 'achievements_counts',
50-
'fields' => [
51-
[
52-
'name' => 'achievement_id',
53-
'type' => 'INT',
54-
'length' => '10',
55-
'unsigned' => true,
56-
'null' => false,
57-
],
58-
[
59-
'name' => 'num_players',
60-
'type' => 'INT',
61-
'length' => '10',
62-
'unsigned' => true,
63-
'null' => false,
64-
],
65-
[
66-
'name' => 'date_updated',
67-
'type' => 'INT',
68-
'length' => '10',
69-
'unsigned' => true,
70-
'null' => false,
71-
],
72-
],
73-
'indexes' => [
74-
[
75-
'type' => 'PRIMARY',
76-
'cols' => [
77-
[
78-
'name' => 'achievement_id',
79-
],
80-
],
81-
],
82-
],
83-
'props' => [
84-
'ENGINE' => 'InnoDB',
85-
'CHARSET' => 'utf8',
86-
],
87-
],
88-
]
43+
```SQL
44+
CREATE TABLE `achievements_counts` (
45+
`achievement_id` int(10) unsigned NOT NULL,
46+
`num_players` int(10) unsigned NOT NULL,
47+
`date_updated` int(10) unsigned NOT NULL,
48+
PRIMARY KEY (`achievement_id`)
49+
) ENGINE=InnoDB DEFAULT CHARSET=utf8;
50+
```
51+
52+
```php
53+
[
54+
'achievements_counts' => [
55+
'name' => 'achievements_counts',
56+
'fields' => [
57+
[
58+
'name' => 'achievement_id',
59+
'type' => 'INT',
60+
'length' => '10',
61+
'unsigned' => true,
62+
'null' => false,
63+
],
64+
[
65+
'name' => 'num_players',
66+
'type' => 'INT',
67+
'length' => '10',
68+
'unsigned' => true,
69+
'null' => false,
70+
],
71+
[
72+
'name' => 'date_updated',
73+
'type' => 'INT',
74+
'length' => '10',
75+
'unsigned' => true,
76+
'null' => false,
77+
],
78+
],
79+
'indexes' => [
80+
[
81+
'type' => 'PRIMARY',
82+
'cols' => [
83+
[
84+
'name' => 'achievement_id',
85+
],
86+
],
87+
],
88+
],
89+
'props' => [
90+
'ENGINE' => 'InnoDB',
91+
'CHARSET' => 'utf8',
92+
],
93+
],
94+
]
95+
```
8996

9097
You can also use the lexer directly to work with other piece of SQL:
9198

92-
$parser = new SQLParser();
93-
$parser->lex($sql);
99+
```php
100+
$parser = new SQLParser();
101+
$parser->lex($sql);
94102

95-
print($parser->tokens);
103+
print($parser->tokens);
104+
```
96105

97-
The `tokens` property contains an array of tokens. SQL keywords are returned as uppercase,
106+
The `tokens` property contains an array of tokens. SQL keywords are returned as uppercase,
98107
with multi-word terms (e.g. `DEFAULT CHARACTER SET`) as a single token. Strings and escaped
99108
identifiers are not further processed; they are returned exactly as expressed in the input SQL.
100109

101-
By default, the tokenizer will ignore unterminated comments and strings, and stop parsing at
110+
By default, the tokenizer will ignore unterminated comments and strings, and stop parsing at
102111
that point, producing no further tokens. You can set `$parser->throw_on_bad_syntax = true;` to
103112
throw an exception of type `iamcal\SQLParserSyntaxException` instead.
104113

105-
106114
## Performance
107115

108116
My test target is an 88K SQL file containing 114 tables from Glitch's main database.
@@ -113,18 +121,16 @@ seconds just to lex the input. This was obviously not a great option.
113121
The current implementation uses a hand-written lexer which takes around 140ms to lex the same
114122
input and imposes less odd restrictions. This seems to be the way to go.
115123

116-
117124
## History
118125

119126
This library was created to parse multiple `CREATE TABLE` schemas and compare them, so
120127
figure out what needs to be done to migrate one to the other.
121128

122129
This is based on the system used at b3ta, Flickr and then Tiny Speck to check the differences
123-
between production and development databases and between shard instances. The original system
130+
between production and development databases and between shard instances. The original system
124131
just showed a diff (see [SchemaDiff](https://github.com/iamcal/SchemaDiff)), but that was a bit
125132
of a pain.
126133

127-
128134
## Unsupported features
129135

130136
MySQL table definitions have a *lot* of options, so some things just aren't supported. They include:
@@ -141,20 +147,19 @@ MySQL table definitions have a *lot* of options, so some things just aren't supp
141147
If you need support for one of these features, open an issue or (better) send a pull request with tests.
142148

143149
The specs for each of the four field groupings can be found here:
150+
144151
* https://dev.mysql.com/doc/refman/5.7/en/numeric-type-syntax.html
145152
* https://dev.mysql.com/doc/refman/5.7/en/date-and-time-type-syntax.html
146153
* https://dev.mysql.com/doc/refman/5.7/en/string-type-syntax.html
147154
* https://dev.mysql.com/doc/refman/5.7/en/spatial-type-overview.html
148155

149-
150156
## Alternatives
151157

152158
If you're using PHP, then [Modyllic](https://github.com/onlinebuddies/modyllic) is a great SQL parser and set of schema management tools.
153159

154160
If you're using Hack, then [Hack SQL Fake](https://github.com/slackhq/hack-sql-fake) allows you to parse SQL and create a fake MySQL
155161
server for testing, with many (but not all!) features of MySQL.
156162

157-
158163
## Publishing
159164

160165
To publish a new version:

0 commit comments

Comments
 (0)