美文网首页
Mysql索引实践

Mysql索引实践

作者: 缄默的石头 | 来源:发表于2017-08-06 13:29 被阅读31次

    Mysql索引实践
    建表语句如下:

    CREATE TABLE `al_dzfp_status` (
      `invoice_status_id` char(21) NOT NULL COMMENT '主键',
      `fpqqlsh` char(21) DEFAULT NULL COMMENT '请求流水号',
      `content` text COMMENT '请求报文',
      `sys_status` varchar(20) DEFAULT NULL,
      `update_time` datetime DEFAULT NULL,
      PRIMARY KEY (`invoice_status_id`),
      KEY `fpqqlsh_index` (`fpqqlsh`) USING BTREE,
      KEY `date_index` (`update_time`) USING BTREE,
      KEY `status_index` (`sys_status`) USING BTREE
    ) ENGINE=InnoDB DEFAULT CHARSET=utf8 ROW_FORMAT=DYNAMIC COMMENT='开票状态表';
    

    表中数据量大约有三万条
    SQL如下:

    -- select count(1) from al_dzfp_status;
    -- 不能使用索引
    EXPLAIN select * from al_dzfp_status where sys_status != 'PUSH_SUCCESS';
    
    -- 0.031s 可使用索引
    explain select * from al_dzfp_status where sys_status = 'ISSUE_FAIL' OR sys_status = 'RECIEVE_SUCCESS';
    
    -- 0.034s 可使用索引
    explain select * from al_dzfp_status where sys_status in ('ISSUE_FAIL' ,'RECIEVE_SUCCESS');
    
    -- 0.052s 可使用索引
    explain select * from al_dzfp_status where sys_status = 'ISSUE_FAIL' UNION select * from al_dzfp_status where sys_status = 'RECIEVE_SUCCESS';
    
    -- 子查询中能使用到索引(type 为ref),where条件是否做索引,视查询出来的结果集大小而定,如果结果集太大也命中不了索引,结果集数量小的话可以走索引
    explain select * from al_dzfp_status where fpqqlsh in (select fpqqlsh from al_dzfp_status where sys_status = 'PUSH_SUCCESS');
    explain select * from al_dzfp_status where fpqqlsh in (select fpqqlsh from al_dzfp_status where sys_status = 'RECIEVE_SUCCESS');
    
    -- 走status的索引,不走update_time索引;使用(status,update_time)联合索引即可
    explain select * from al_dzfp_status where sys_status in ('ISSUE_FAIL' ,'RECIEVE_SUCCESS') order by update_time DESC;
    
    -- 不走索引
    EXPLAIN select * from al_dzfp_status where sys_status = 'ISSUE_FAIL' or update_time < CURDATE();
    

    相关文章

      网友评论

          本文标题:Mysql索引实践

          本文链接:https://www.haomeiwen.com/subject/acwglxtx.html