在Percona Performance Conference 2009大会上来自yahoo的Surat Singh Bhati (surat@yahoo-inc.com) 和 Rick James (rjames@yahoo-inc.com)给大家分享了MySQL高效分页的经验。
[size=large][b]一、概述[/b][/size]
[list]
[*]常见分页方式
[*]schema设计和常见的分页方式(偏移)
[*]避免分页偏移过大的技巧
[*]性能对比
[*]重点
[/list]
[size=large][b]二、常见分页方式[/b][/size]
[img]/upload/attachment/122203/d161d9bf-45fc-3f13-acc1-be5eba13bbf2.jpg[/img]
[size=large][b]三.前提[/b][/size]
大记录表要高效分页
[list]
[*]WHERE条件使用索引完成
[*]WHERE条件和排序能够使用同个索引完成
[/list]基础知识[list]
[*][url]http://dev.mysql.com/doc/refman/5.1/en/mysql-indexes.html[/url]
[*][url]http://dev.mysql.com/doc/refman/5.1/en/order-by-optimization.html[/url]
[*][url]http://dev.mysql.com/doc/refman/5.1/en/limit-optimization.html[/url]
[/list]
索引 a_b_c (a, b, c)
下面的查询可以使用索引来解决ORDER部分：
[list]
[*]ORDER BY a
[*]ORDER BY a,b
[*]ORDER BY a, b, c
[*]ORDER BY a DESC, b DESC, c DESC
[/list]
下面的查询可以使用索引来解决WHERE和ORDER部分：:
[list]
[*]WHERE a = const ORDER BY b, c
[*]WHERE a = const AND b = const ORDER BY c
[*]WHERE a = const ORDER BY b, c
[*]WHERE a = const AND b > const ORDER BY b, c
[/list]
下面的查询无法使用索引完成，需额外排序：
[list]
[*]ORDER BY a ASC, b DESC, c DESC /* 混合ASC和DESC */
[*]WHERE g = const ORDER BY b, c /* 字段g不是索引一部分 */
[*]WHERE a = const ORDER BY c /* 没有使用字段b */
[*]WHERE a = const ORDER BY a, d /* 字段d不是索引的一部分 */
[/list]
[size=large][b]四、Schema 设计[/b][/size]
CREATE TABLE `message` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`title` varchar(255) COLLATE utf8_unicode_ci NOT NULL,
`user_id` int(11) NOT NULL,
`content` text COLLATE utf8_unicode_ci NOT NULL,
`create_time` int(11) NOT NULL,
`thumbs_up` int(11) NOT NULL DEFAULT '0', /* 投票数 */
PRIMARY KEY (`id`),
KEY `thumbs_up_key` (`thumbs_up`,`id`)
) ENGINE=InnoDB
mysql> show table status like 'message' \G
Engine: InnoDB
Version: 10
Row_format: Compact
Rows: 50000040 /* 5千万 */
Avg_row_length: 565
Data_length: 28273803264 /* 26 GB */
Index_length: 789577728 /* 753 MB */
Data_free: 6291456
Create_time: 2009-04-20 13:30:45
两个分页例子：[list]
[*]按照time(发布时间)分页，新发布的在前面
[*]按照thumps_up(投票数)分页，票高的在前面
[/list]
[size=large][b]五、典型的分页查询[/b][/size]
[b]1.统计记录数量[/b]
SELECT count(*) FROM message
[b]2. 查询当前页[/b]
SELECT * FROM message ORDER BY id DESC LIMIT 0, 20
[list]
[*]http://domain.com/message?page=1
ORDER BY id DESC LIMIT 0, 20
[*]http://domain.com/message?page=2
ORDER BY id DESC LIMIT 20, 20
[*]http://domain.com/message?page=3
ORDER BY id DESC LIMIT 40, 20
[/list]
提示：id 是自动增长的(auto_increment)，通过id就可以取得最新的列表，不需要创建专门记录时间的字段。
[size=large][b]六、explain[/b][/size]
mysql> explain SELECT * FROM message
ORDER BY id DESC
LIMIT 10000, 20\G
***************** 1. row **************
id: 1
select_type: SIMPLE
table: message
type: index
possible_keys: NULL
key: PRIMARY
key_len: 4
ref: NULL
rows: 10020
Extra:
1 row in set (0.00 sec)
[list]
[*]它可以使用索引，并且只要找到需要的结果后就停止扫描.
[*]LIMIT 10000, 20 需要读取前10000行，然后获取后面的20行
[/list]
[size=large][b]六、瓶颈[/b][/size]
[list]
[*]较大的偏移(OFFSET)会增加结果集， MySQL has to bring data in memory that is never returned to caller.
[*]Performance issue is more visible when your have database that can't fit in main memory.
[*]小比例的低效分页足够产生磁盘I/O瓶颈
[*]为了显示“第 21条 至 40条 (共 1000000)，需要统计1000000行
[/list]
[size=large][b]七、简单的解决方法[/b][/size]
[list=1]
[*]不显示记录总数，没用户在乎这个数字
[*]不让用户访问页数比较大的记录，重定向他们
[/list]
[size=large][b]八、避免count(*)[/b][/size]
[list=1]
[*]不显示总数，让用户通过“下一页”来翻页
[*]缓存总数，显示一个大概值，没有用户在乎是324533条还是324633 （译：测试在乎-_-!!）
[*]Display 41 to 80 of Thousands
[*]单独统计总数，在插入和删除时递增/递减
[/list]
[size=large][b]九、解决偏移查询[/b][/size]
[list=1]
[*]更改ui，不提供跳到某页的按钮
[*]LIMIT N 是高效的, 但不要使用 LIMIT M,N
[list]
[*]从WHERE条件里找到分页(LIMIT N)的线索
[*]Find the desired records using more restricted WHERE using given clue and ORDER BY and LIMIT N without OFFSET)
[/list]
[/list]
[size=large][b]十、寻找线索[/b][/size]
[img]/upload/attachment/122318/7d3930ab-5602-3245-a651-5f35cf91d1a7.jpg[/img]
译：last_seen是id。这里的分页只有“上一页”、“下一页” 按钮
[size=large][b]十一、根据线索解决方案[/b][/size]
下一页:
http://domain.com/forum?page=2&last_seen=100&dir=next
WHERE id < 100 /* last_seen */
ORDER BY id DESC LIMIT $page_size /* 没有偏移 */
上一页:
http://domain.com/forum?page=1&last_seen=98&dir=prev
WHERE id > 98 /* last_seen */
ORDER BY id ASC LIMIT $page_size /* 没有偏移 */
译：通过每页第一条或最后一条记录的id来做条件筛选，再配合降序和升序获得上/下一页的结果集
[size=large][b]十二、根据线索解决方案[/b][/size]
mysql> explain
SELECT * FROM message
WHERE id < '49999961'
ORDER BY id DESC LIMIT 20 \G
*************************** 1. row ***************************
id: 1
select_type: SIMPLE
table: message
type: range
possible_keys: PRIMARY
key: PRIMARY
key_len: 4
ref: NULL
Rows: 25000020 /* 忽略这里 */
Extra: Using where
1 row in set (0.00 sec)
[size=large][b]十三、当你排序的字段不是唯一的，怎么办?[/b][/size]
[quote]99
99
98 第一页
98
98
98
98
97 第二页
97
10[/quote]
我们不能这样查询:
WHERE thumbs_up < 98
ORDER BY thumbs_up DESC /* 结果将返回重复的记录 */
我们可以这样查询:
WHERE thumbs_up <= 98
AND <额外的条件>
ORDER BY thumbs_up DESC
[size=large][b]十四、额外的条件[/b][/size]
[list]
[*]考虑到 thumbs_up 是“主要字段”，如果我们添加一个“次要字段”，我们可以使用“主要字段”和“次要字段”作为查询条件
[*]其次，我们可以考虑使用id(primary key)作为我们的次要字段
[/list]
[size=large][b]十五、解决方案[/b][/size]
第一页：
SELECT thumbs_up, id
FROM message
ORDER BY thumbs_up DESC, id DESC
LIMIT $page_size
+-----------+----+
| thumbs_up | id |
+-----------+----+
| 99 | 14 |
| 99 | 2 |
| 98 | 18 |
| 98 | 15 |
| 98 | 13 |
+-----------+----+
下一页：
SELECT thumbs_up, id
FROM message
WHERE thumbs_up <= 98 AND (id < 13 OR thumbs_up < 98)
ORDER BY thumbs_up DESC, id DESC
LIMIT $page_size
+-----------+----+
| thumbs_up | id |
+-----------+----+
| 98 | 10 |
| 98 | 6 |
| 97 | 17 |
[size=large][b]十六、优化[/b][/size]
查询:
SELECT * FROM message
WHERE thumbs_up <= 98
AND (id < 13 OR thumbs_up < 98)
ORDER BY thumbs_up DESC, id DESC
LIMIT 20
我们可以这样写:
SELECT m2.* FROM message m1, message m2
WHERE m1.id = m2.id
AND m1.thumbs_up <= 98
AND (m1.id < 13 OR m1.thumbs_up < 98)
ORDER BY m1.thumbs_up DESC, m1.id DESC
LIMIT 20;
[size=large][b]十七、explain[/b][/size]
*************************** 1. row ***************************
id: 1
select_type: SIMPLE
table: m1
type: range
possible_keys: PRIMARY,thumbs_up_key
key: thumbs_up_key /* (thumbs_up,id) */
key_len: 4
ref: NULL
Rows: 25000020 /* 忽略这里 */
Extra: Using where; Using index /* Cover 译：Cover就是说所需要的数据之从索引里获取就可以满足了 */
*************************** 2. row ***************************
id: 1
select_type: SIMPLE
table: m2
type: eq_ref
possible_keys: PRIMARY
key: PRIMARY
key_len: 4
ref: forum.m1.id
rows: 1
Extra:
[size=large][b]十八、性能提升[/b][/size]
[img]/upload/attachment/124012/447f5046-3d39-3a9d-8f7c-72f123e9a8d4.jpg[/img]
[img]/upload/attachment/124010/f4569ac9-cada-386f-83e8-6b4cbdd57a22.jpg[/img]
[size=large][b]十九、吞吐量提升[/b][/size]
每页30条记录，查看第一页的话，使用 LIMIT OFFSET, N方式，可以达到 600 次查询/秒，如果使用 LIMIT N (无偏移)方式，提升到 3.7k 次查询/秒
[size=large][b]二十、Bonus Point[/b][/size]
Product issue with LIMIT M, N
User is reading a page, in the mean time some records may be added to
previous page.
Due to insert/delete pages records are going to move forward/backward
as rolling window:
– User is reading messages on 4th page
– While he was reading, one new message posted (it would be there on page
one), all pages are going to move one message to next page.
– User Clicks on Page 5
– One message from page got pushed forward on page 5, user has to read it
again
No such issue with news approach
[size=large][b]二十一、不足[/b][/size]
SEO专家会说:Let bot reach all you pages with fewer number of deep dive
[b]两个解决方案：[/b]
Two Solutions:
• Read extra rows
– Read extra rows in advance and construct links for few previous & next pages
• Use small offset
– Do not read extra rows in advance, just add links for few past & next pages
with required offset & last_seen_id on current page
– Do query using new approach with small offset to display desired page
[img]http://dl.iteye.com/upload/attachment/163016/26f3f5f8-11eb-3bf3-95b5-34fe6fd1c63b.png[/img]
Additional concern: Dynamic urls, last_seen is not constant over time.
原文：[url=/admin/blogs/]http://www.percona.com/ppc2009/PPC2009_mysql_pagination.pdf[/url]
本文内容由网友自发贡献，版权归原作者所有，本站不承担相应法律责任。如您发现有涉嫌抄袭侵权的内容，请联系:hwhale#tublm.com(使用前将#替换为@)
