Question

我在视图中使用了以下查询：

select `a`.`device_id` AS `device_id`,
       `a`.`alias` AS `alias`,
       `a`.`freq` AS `freq`,
       `a`.`gateway` AS `gateway`,
       `a`.`device_lat` AS`device_lat`,
       `a`.`device_long` AS `device_long`,
       `a`.`device_disabled` AS `device_disabled`,
       count(`b`.`msg_id`) AS `total_messages`,
       avg(`b`.`rssi`) AS `avg_rssi`,
       max(`b`.`db_timestamp`) AS `last_active`,
       (now() <= (max(`b`.`db_timestamp`) + interval 3 hour)) AS `device_status`
from `demo`.`lora_device` `a` 
left join `demo`.`lora_message` `b` on `a`.`device_id` = `b`.`eui` 
group by `a`.`device_id`

此查询大约需要4秒钟加载，有关标签的信息：

lora_message：25k行约20列

lora_device：520行约10列

通常我会说这对mysql来说不是问题，但由于某种原因它会变得很慢。

Answer 1

尝试添加索引

 create index ix_loramessage_rssi on lora_message(eui, rssi)
 create index ix_loramessage_db_timestamp on lora_message(eui, db_timestamp)

并使用

 count(`b`.`rssi`) AS `total_messages`,

而不是

 count(`b`.`msg_id`) AS `total_messages`,

因为它应该在您的查询中返回相同的结果

MySQL查询查看速度慢的表现

1 个答案: