MySQL查询计划
1、Explain显示查询计划
explain ops stmt，stmt可以是select/insert/update/delete，其中ops如下
（1）EXTENDED
显示SQL语句的详细查询执行计划，之后可以通过SHOW WARNINGS命令查看详细的信息
（2）PARTITIONS
显示SQL语句的带有分区表信息的查询执行计划
（3）FORMAT = TRADITION/JSON
TRADITIION：传统类型，按行隔离
JSON：JSON格式
2、查询实例
（1）建表
CREATE TABLE t1 (id1 INT, a1 INT, b1 INT, PRIMARY KEY(id1)); CREATE TABLE t2 (id2 INT, a2 INT, b2 INT); CREATE TABLE t3 (id3 INT UNIQUE, a3 INT, b3 INT); CREATE TABLE t4 (id4 INT, a4 INT, b4 INT); CREATE TABLE t5 (id5 INT UNIQUE, a5 INT, b5 INT);
（2）准备数据
准备t1
DELIMITER $$ CREATE DEFINER=`root`@`localhost` PROCEDURE `insertT1`() BEGIN set @counter=0; while @counter < 10000 do SET @col2 = FLOOR(1 + (RAND() * 10000)); set @col3 = FLOOR(1 + (RAND() * 10000)); insert into t1(id1,a1,b1) values(@counter,@col2,@col3); set @counter = @counter +1; END while; END$$ DELIMITER ;
call insertT1;
准备t2
DELIMITER $$ CREATE DEFINER=`root`@`localhost` PROCEDURE `insertT2`() BEGIN SET @counter = 0; while @counter < 100 do SET @col2 = FLOOR(1 + (RAND() * 10000)); set @col3 = FLOOR(1 + (RAND() * 10000)); insert into t2(id2,a2,b2) values(@counter,@col2,@col3); set @counter = @counter +1; end while; END$$ DELIMITER ;
准备t3
DELIMITER $$ CREATE DEFINER=`root`@`localhost` PROCEDURE `insertT3`() BEGIN SET @counter = 0; while @counter < 100 do SET @col2 = FLOOR(1 + (RAND() * 10000)); set @col3 = FLOOR(1 + (RAND() * 10000)); insert into t3(id3,a3,b3) values(@counter,@col2,@col3); set @counter = @counter +1; end while; END$$ DELIMITER ;
准备t4
DELIMITER $$ CREATE DEFINER=`root`@`localhost` PROCEDURE `insertT4`() BEGIN SET @counter = 0; while @counter < 7 do SET @col2 = FLOOR(1 + (RAND() * 10000)); set @col3 = FLOOR(1 + (RAND() * 10000)); insert into t4(id4,a4,b4) values(@counter,@col2,@col3); set @counter = @counter +1; end while; END$$ DELIMITER ;
准备t5
DELIMITER $$ CREATE DEFINER=`root`@`localhost` PROCEDURE `insertT5`() BEGIN SET @counter = 0; while @counter < 10 do SET @col2 = FLOOR(1 + (RAND() * 10000)); set @col3 = FLOOR(1 + (RAND() * 10000)); insert into t5(id5,a5,b5) values(@counter,@col2,@col3); set @counter = @counter +1; end while; END$$ DELIMITER ;
（5）explain
explain select * from (t1 left join t2 on true),(t3 full join t4 on true),t5 where id1=id2 and id2 = id3 and id3 = id4 and id4 = id5;
结果
说明：
A、id表示对象操作顺序，数字越大，越先被执行
B、表连接顺序为t4，t5，t3，t1，t2，因为t4表的记录最少，因此mysql优化后，先连接t4
C、因为t5，t3，t1有索引可以用，因此用索引定位这三个表的数据
D、t2表的数据多，而且又没有索引可以用，因此放到最后连接。连接使用了Extra列表明的块嵌套循环连接算法，并且使用了连接缓存。