【原创】MySQL 5.7 SYS系统SCHEMA

在说明系统数据库之前，先来看下MySQL在数据字典方面的演变历史：
MySQL4.1 提供了information_schema 数据字典。从此可以很简单的用SQL语句来检索需要的系统元数据了。
MySQL5.5 提供了performance_schema 性能字典。 但是这个字典比较专业，一般人可能也就看看就不了了之了。
MySQL5.7 提供了 sys系统数据库。 sys数据库里面包含了一系列的存储过程、自定义函数以及视图来帮助我们快速的了解系统的元数据信息。

sys系统数据库结合了information_schema和performance_schema的相关数据，让我们更加容易的检索元数据。 现在呢，我就示范下几种场景下如何快速的使用。

第一，
比如之前想要知道某个表是否存在与否，可以用以下两种方法：
A， 悲观的方法，写SQL从information_schema中拿信息：

mysql> SELECT IF(COUNT(*) = 0,'Not exists!','Exists!') AS 'result' FROM information_schema.tables WHERE table_schema = 'new_feature' AND table_name = 't1';
+-------------+
| result      |
+-------------+
| Not exists! |
+-------------+
1 row in set (0.00 sec)

B，乐观的方法，假设表存在，写一个存储过程：

DELIMITER $$
USE `new_feature`$$
DROP PROCEDURE IF EXISTS `sp_table_exists`$$
CREATE DEFINER=`root`@`localhost` PROCEDURE `sp_table_exists`(
IN db_name VARCHAR(64),
IN tb_name VARCHAR(64),
OUT is_exists VARCHAR(60)
)
BEGIN
DECLARE no_such_table CONDITION FOR 1146;
DECLARE EXIT HANDLER FOR no_such_table
BEGIN
SET is_exists = 'Not exists!';
END;

SET @stmt = CONCAT('select 1 from ',db_name,'.',tb_name);
PREPARE s1 FROM @stmt;
EXECUTE s1;
DEALLOCATE PREPARE s1;
SET is_exists = 'Exists!';
END$$
DELIMITER ;

现在来调用：

mysql> call sp_table_exists('new_feature','t1',@result);
Query OK, 0 rows affected (0.00 sec)
mysql> select @result;
+-------------+
| @result     |
+-------------+
| Not exists! |
+-------------+
1 row in set (0.00 sec)

现在我们直接用sys数据库里面现有的存储过程来进行调用，

mysql> CALL table_exists('new_feature','t1',@v_is_exists);
Query OK, 0 rows affected (0.00 sec)
mysql> SELECT IF(@v_is_exists = '','Not exists!',@v_is_exists) AS 'result';
+-------------+
| result      |
+-------------+
| Not exists! |
+-------------+
1 row in set (0.00 sec)

第二，获取没有使用过的索引。

mysql> SELECT * FROM schema_unused_indexes;
+---------------+-------------+--------------+
| object_schema | object_name | index_name   |
+---------------+-------------+--------------+
| new_feature   | t1          | idx_log_time |
| new_feature   | t1          | idx_rank2    |
+---------------+-------------+--------------+
2 rows in set (0.00 sec)

第三， 检索指定数据库下面的表扫描信息，过滤出执行次数大于10的查询，

mysql> SELECT * FROM statement_analysis WHERE db='new_feature' AND full_scan = '*'  AND exec_count > 10\G
*************************** 1. row ***************************
query: SHOW STATUS
db: new_feature
full_scan: *
exec_count: 26
err_count: 0
warn_count: 0
total_latency: 74.68 ms
max_latency: 3.86 ms
avg_latency: 2.87 ms
lock_latency: 4.50 ms
rows_sent: 9594
rows_sent_avg: 369
rows_examined: 9594
rows_examined_avg: 369
rows_affected: 0
rows_affected_avg: 0
tmp_tables: 0
tmp_disk_tables: 0
rows_sorted: 0
sort_merge_passes: 0
digest: 475fa3ad9d4a846cfa96441050fc9787
first_seen: 2015-11-16 10:51:17
last_seen: 2015-11-16 11:28:13
*************************** 2. row ***************************
query: SELECT `state` , `round` ( SUM ... uration (summed) in sec` DESC
db: new_feature
full_scan: *
exec_count: 12
err_count: 0
warn_count: 12
total_latency: 16.43 ms
max_latency: 2.39 ms
avg_latency: 1.37 ms
lock_latency: 3.54 ms
rows_sent: 140
rows_sent_avg: 12
rows_examined: 852
rows_examined_avg: 71
rows_affected: 0
rows_affected_avg: 0
tmp_tables: 24
tmp_disk_tables: 0
rows_sorted: 140
sort_merge_passes: 0
digest: 538e506ee0075e040b076f810ccb5f5c
first_seen: 2015-11-16 10:51:17
last_seen: 2015-11-16 11:28:13
2 rows in set (0.01 sec)

第四， 同样继续上面的，过滤出有临时表的查询，

mysql> SELECT * FROM statement_analysis WHERE db='new_feature' AND tmp_tables > 0 ORDER BY tmp_tables DESC LIMIT 1\G
*************************** 1. row ***************************
query: SELECT `performance_schema` .  ... name` . `SUM_TIMER_WAIT` DESC
db: new_feature
full_scan: *
exec_count: 2
err_count: 0
warn_count: 0
total_latency: 87.96 ms
max_latency: 59.50 ms
avg_latency: 43.98 ms
lock_latency: 548.00 us
rows_sent: 101
rows_sent_avg: 51
rows_examined: 201
rows_examined_avg: 101
rows_affected: 0
rows_affected_avg: 0
tmp_tables: 332
tmp_disk_tables: 15
rows_sorted: 0
sort_merge_passes: 0
digest: ff9bdfb7cf3f44b2da4c52dcde7a7352
first_seen: 2015-11-16 10:24:42
last_seen: 2015-11-16 10:24:42
1 row in set (0.01 sec)

可以看到上面查询详细的详细，再也不用执行show status 手工去过滤了。

第五， 检索执行次数排名前五的语句，

mysql> SELECT statement,total FROM user_summary_by_statement_type WHERE `user`='root' ORDER BY total DESC LIMIT 5;
+-------------------+-------+
| statement         | total |
+-------------------+-------+
| jump_if_not       | 17635 |
| freturn           |  3120 |
| show_create_table |   289 |
| Field List        |   202 |
| set_option        |   190 |
+-------------------+-------+
5 rows in set (0.01 sec)

示例我就写这么多了，详细的去看使用手册并且自己摸索去吧。

本文出自 “上帝，咱们不见不散！” 博客，请务必保留此出处http://yueliangdao0608.blog.51cto.com/397025/1716616