OPTIMIZE TABLE linked list 表优化原理 链表数据结构 空间再利用

 小结：

1、加快读写；

2、对于InnoDB表，在一定条件下，通过复制旧表重建；

3、实践中，

3.1、show processlist;查看线程，发现，认为堵塞读请求；

3.2、数据长度空间不变，索引长度空间减小；历史的写操作是：url访问统计表，当天多次更新当天聚合结果数据的逻辑是“当天数据删除，再插入”（url-pv）；

MySQL :: MySQL 8.0 Reference Manual :: 13.7.3.4 OPTIMIZE TABLE Syntax https://dev.mysql.com/doc/refman/8.0/en/optimize-table.html

MySQL 8.0 Reference Manual  /  ...  /  OPTIMIZE TABLE Syntax

13.7.3.4 OPTIMIZE TABLE Syntax

OPTIMIZE [NO_WRITE_TO_BINLOG | LOCAL]
    TABLE tbl_name [, tbl_name] ...

OPTIMIZE TABLE reorganizes the physical storage of table data and associated index data, to reduce storage space and improve I/O efficiency when accessing the table. The exact changes made to each table depend on the storage engine used by that table.

Use OPTIMIZE TABLE in these cases, depending on the type of table:

• After doing substantial insert, update, or delete operations on an InnoDB table that has its own .ibd file because it was created with theinnodb_file_per_table option enabled. The table and indexes are reorganized, and disk space can be reclaimed for use by the operating system.

• After doing substantial insert, update, or delete operations on columns that are part of a FULLTEXT index in an InnoDB table. Set the configuration option innodb_optimize_fulltext_only=1 first. To keep the index maintenance period to a reasonable time, set theinnodb_ft_num_word_optimize option to specify how many words to update in the search index, and run a sequence of OPTIMIZE TABLEstatements until the search index is fully updated.

• After deleting a large part of a MyISAM or ARCHIVE table, or making many changes to a MyISAM or ARCHIVE table with variable-length rows (tables that have VARCHAR, VARBINARY, BLOB, or TEXT columns). Deleted rows are maintained in a linked list and subsequent INSERToperations reuse old row positions. You can use OPTIMIZE TABLE to reclaim the unused space and to defragment the data file. After extensive changes to a table, this statement may also improve performance of statements that use the table, sometimes significantly.

This statement requires SELECT and INSERT privileges for the table.

OPTIMIZE TABLE works for InnoDB, MyISAM, and ARCHIVE tables. OPTIMIZE TABLE is also supported for dynamic columns of in-memory NDBtables. It does not work for fixed-width columns of in-memory tables, nor does it work for Disk Data tables. The performance of OPTIMIZE on NDB Cluster tables can be tuned using --ndb_optimization_delay, which controls the length of time to wait between processing batches of rows by OPTIMIZE TABLE. For more information, see Previous NDB Cluster Issues Resolved in NDB Cluster 7.3.

For NDB Cluster tables, OPTIMIZE TABLE can be interrupted by (for example) killing the SQL thread performing the OPTIMIZE operation.

By default, OPTIMIZE TABLE does not work for tables created using any other storage engine and returns a result indicating this lack of support. You can make OPTIMIZE TABLE work for other storage engines by starting mysqld with the --skip-new option. In this case, OPTIMIZE TABLE is just mapped to ALTER TABLE.

This statement does not work with views.

OPTIMIZE TABLE is supported for partitioned tables. For information about using this statement with partitioned tables and table partitions, see Section 23.3.4, “Maintenance of Partitions”.

By default, the server writes OPTIMIZE TABLE statements to the binary log so that they replicate to replication slaves. To suppress logging, specify the optional NO_WRITE_TO_BINLOG keyword or its alias LOCAL.

• OPTIMIZE TABLE Output

• InnoDB Details

• MyISAM Details

• Other Considerations

OPTIMIZE TABLE Output

OPTIMIZE TABLE returns a result set with the columns shown in the following table.

Column	Value
Table	The table name
Op	Always optimize
Msg_type	status, error, info, note, or warning
Msg_text	An informational message

OPTIMIZE TABLE table catches and throws any errors that occur while copying table statistics from the old file to the newly created file. For example. if the user ID of the owner of the .MYD or .MYI file is different from the user ID of the mysqld process, OPTIMIZE TABLE generates a "cannot change ownership of the file" error unless mysqld is started by the root user.

InnoDB Details

For InnoDB tables, OPTIMIZE TABLE is mapped to ALTER TABLE ... FORCE, which rebuilds the table to update index statistics and free unused space in the clustered index. This is displayed in the output of OPTIMIZE TABLE when you run it on an InnoDB table, as shown here:

mysql> OPTIMIZE TABLE foo;
+----------+----------+----------+-------------------------------------------------------------------+
| Table    | Op       | Msg_type | Msg_text                                                          |
+----------+----------+----------+-------------------------------------------------------------------+
| test.foo | optimize | note     | Table does not support optimize, doing recreate + analyze instead |
| test.foo | optimize | status   | OK                                                                |
+----------+----------+----------+-------------------------------------------------------------------+

OPTIMIZE TABLE uses online DDL for regular and partitioned InnoDB tables, which reduces downtime for concurrent DML operations. The table rebuild triggered by OPTIMIZE TABLE and performed under the cover by ALTER TABLE ... FORCE is completed in place. An exclusive table lock is only taken briefly during the prepare phase and the commit phase of the operation. During the prepare phase, metadata is updated and an intermediate table is created. During the commit phase, table metadata changes are committed.

OPTIMIZE TABLE rebuilds the table using the table copy method under the following conditions:

• When the old_alter_table system variable is enabled.

• When the mysqld --skip-new option is enabled.

OPTIMIZE TABLE using online DDL is not supported for InnoDB tables that contain FULLTEXT indexes. The table copy method is used instead.

InnoDB stores data using a page-allocation method and does not suffer from fragmentation in the same way that legacy storage engines (such as MyISAM) will. When considering whether or not to run optimize, consider the workload of transactions that your server will process:

• Some level of fragmentation is expected. InnoDB only fills pages 93% full, to leave room for updates without having to split pages.

• Delete operations might leave gaps that leave pages less filled than desired, which could make it worthwhile to optimize the table.

• Updates to rows usually rewrite the data within the same page, depending on the data type and row format, when sufficient space is available. See Section 15.9.1.5, “How Compression Works for InnoDB Tables” and Section 15.10, “InnoDB Row Formats”.

• High-concurrency workloads might leave gaps in indexes over time, as InnoDB retains multiple versions of the same data due through itsMVCC mechanism. See Section 15.3, “InnoDB Multi-Versioning”.

MyISAM Details

For MyISAM tables, OPTIMIZE TABLE works as follows:

1. If the table has deleted or split rows, repair the table.

2. If the index pages are not sorted, sort them.

3. If the table's statistics are not up to date (and the repair could not be accomplished by sorting the index), update them.

Other Considerations

OPTIMIZE TABLE is performed online for regular and partitioned InnoDB tables. Otherwise, MySQL locks the table during the time OPTIMIZE TABLE is running.

OPTIMIZE TABLE does not sort R-tree indexes, such as spatial indexes on POINT columns. (Bug #23578)

实践

show processlist; 

statistics_visit Query 1238 altering table OPTIMIZE TABLE `test`
community Sleep 56
community Sleep 20
statistics_visit Sleep 158
statistics_visit Sleep 155
statistics_visit Query 0 init show processlist
statistics_visit Execute 104 Creating sort index SELECT address,pv,modify_time,ip FROM test WHERE uid=50016942 ORDER BY modify_time DESC LIMIT
statistics_visit Execute 104 Sending data SELECT SUM(pv) AS c FROM test WHERE uid=50016942 UNION ALL SELECT SUM(pv) AS c FROM st
statistics_visit Execute 96 Sending data SELECT SUM(pv) AS c FROM test WHERE uid=50016942
statistics_visit Execute 96 Creating sort index SELECT address,pv,modify_time,ip FROM test WHERE uid=50016942 ORDER BY modify_time DESC LIMIT
statistics_visit Execute 41 Sending data SELECT SUM(pv) AS c FROM test WHERE uid=50015592
statistics_visit Execute 40 Creating sort index SELECT address,pv,modify_time,ip FROM test WHERE uid=50015592 ORDER BY modify_time DESC LIMIT
statistics_visit Execute 35 Creating sort index SELECT address,pv,modify_time,ip FROM test WHERE uid=50017425 ORDER BY modify_time DESC LIMIT
statistics_visit Execute 35 Sending data SELECT SUM(pv) AS c FROM test WHERE uid=50017425 UNION ALL SELECT SUM(pv) AS c FROM st
statistics_visit Execute 34 Creating sort index SELECT address,pv,modify_time,ip FROM test WHERE uid=50015592 ORDER BY modify_time DESC LIMIT
statistics_visit Execute 33 Sending data SELECT SUM(pv) AS c FROM test WHERE uid=50015592 UNION ALL SELECT SUM(pv) AS c FROM st
statistics_visit Execute 18 Sending data SELECT SUM(pv) AS c FROM test WHERE uid=50017828
statistics_visit Execute 18 Creating sort index SELECT address,pv,modify_time,ip FROM test WHERE uid=50017828 ORDER BY modify_time DESC LIMIT

 优化前

 

优化后

 

优化结果语句

Table does not support optimize, doing recreate + analyze instead

 在优化过程中，只有读查询；

数据长度空间不变，索引长度空间由7.63GB减少至5.25GB；