Question

我的feed_listingjob有2个日期时间字段：

+------------+-------------+------+-----+---------+----------------+
| Field      | Type        | Null | Key | Default | Extra          |
+------------+-------------+------+-----+---------+----------------+
| id         | int(11)     | NO   | PRI | NULL    | auto_increment |
| data       | longtext    | NO   |     | NULL    |                |
| meta_data  | longtext    | NO   |     | NULL    |                |
| state      | varchar(25) | NO   |     | NULL    |                |
| error      | longtext    | NO   |     | NULL    |                |
| job_id     | int(11)     | NO   | MUL | NULL    |                |
| created_at | datetime(6) | NO   | MUL | NULL    |                |
| updated_at | datetime(6) | NO   | MUL | NULL    |                |
| es_sync_at | datetime(6) | YES  | MUL | NULL    |                |
+------------+-------------+------+-----+---------+----------------+

updated_at和es_sync_at都会单独编制索引，如下所示：

mysql> show indexes from feed_listingjob;
+-----------------+------------+--------------------------------------------------------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| Table           | Non_unique | Key_name                                                     | Seq_in_index | Column_name | Collation | Cardinality | Sub_part | Packed | Null | Index_type | Comment | Index_comment |
+-----------------+------------+--------------------------------------------------------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| feed_listingjob |          0 | PRIMARY                                                      |            1 | id          | A         |       64534 |     NULL | NULL   |      | BTREE      |         |               |
| feed_listingjob |          1 | feed_listingjob_job_id_4c3b1b514481f269_fk_feed_importjob_id |            1 | job_id      | A         |        2081 |     NULL | NULL   |      | BTREE      |         |               |
| feed_listingjob |          1 | feed_listingjob_fde81f11                                     |            1 | created_at  | A         |       64534 |     NULL | NULL   |      | BTREE      |         |               |
| feed_listingjob |          1 | feed_listingjob_afd1a1a8                                     |            1 | updated_at  | A         |       64534 |     NULL | NULL   |      | BTREE      |         |               |
| feed_listingjob |          1 | feed_listingjob_381895a2                                     |            1 | es_sync_at  | A         |           2 |     NULL | NULL   | YES  | BTREE      |         |               |
+-----------------+------------+--------------------------------------------------------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
5 rows in set (0.00 sec)

我的查询没有使用索引：

mysql> explain SELECT `feed_listingjob`.`id` FROM `feed_listingjob` WHERE `feed_listingjob`.`es_sync_at` < `feed_listingjob`.`updated_at` LIMIT 10;
+----+-------------+-----------------+------+---------------+------+---------+------+-------+-------------+
| id | select_type | table           | type | possible_keys | key  | key_len | ref  | rows  | Extra       |
+----+-------------+-----------------+------+---------------+------+---------+------+-------+-------------+
|  1 | SIMPLE      | feed_listingjob | ALL  | NULL          | NULL | NULL    | NULL | 53534 | Using where |
+----+-------------+-----------------+------+---------------+------+---------+------+-------+-------------+
1 row in set (0.01 sec)

有人可以告诉我为什么以及如何优化此查询？

Answer 1

此处无法使用索引，因为feed_listingjob。updated_at不是常量。这意味着必须检查每一行以检查条件。使用索引的一种方法是使用另一列来存储es_sync_at和updated_at时间戳之间的差异。如果将其存储为差值=（es_sync_at - updated_at timestamps），则查询变为（差异<0）。

MySQL性能问题比较同一个表上的2个DateTime字段

1 个答案: