我有以下mySql语句,我相信它可以提高效率,我只是不确定什么是最好的方法...唯一的区别是WHERE子句中更改的equip_id。 ..
==============
(SELECT
`receipt_ts`,
`driver_id`,
`equip_id`,
`pos_lon`,
`pos_lat`,
`pos_timestamp`
FROM `log_messaging`
WHERE `equip_id`='207' AND `tran_type`='T.2.12.0'
ORDER BY receipt_ts DESC LIMIT 1 )
UNION
(SELECT
`receipt_ts`,
`driver_id`,
`equip_id`,
`pos_lon`,
`pos_lat`,
`pos_timestamp`
FROM `log_messaging`
WHERE `equip_id`='212' AND `tran_type`='T.2.12.0'
ORDER BY receipt_ts DESC LIMIT 1 )
UNION
(SELECT
`receipt_ts`,
`driver_id`,
`equip_id`,
`pos_lon`,
`pos_lat`,
`pos_timestamp`
FROM `log_messaging`
WHERE `equip_id`='213' AND `tran_type`='T.2.12.0'
ORDER BY receipt_ts DESC LIMIT 1 );
=====
我正在联合每个返回的行,为多个equip_id建立一个max(receipt_ts)数据集。 (我需要获得最新的设备定位。
有时,查询最终会出现100多个独特的equip_id。
我正在努力使查询比现在更快地执行(对于~100 UNIONS约为7秒......如上所述...
指出我正确的方向?
谢谢!
答案 0 :(得分:1)
我会使用IN
子句:
SELECT receipt_ts, driver_id, equip_id, pos_lon, pos_lat, pos_timestamp
FROM log_messaging a
JOIN (SELECT c.equip_id, max(c.receipt_ts) as receipt
FROM log_messaging c
WHERE equip_id in ('207', '212', '213')
AND tran_type='T.2.12.0'
GROUP by c.equip_id) b USING(equip_id)
WHERE b.receipt = a.receipt_ts
ORDER BY a.receipt_ts
请注意,如果您真的想使用UNION
(我不明白为什么会这样做)但想要优化它,您可以使用UNION ALL
作为{UNION
1}}检查重复数据删除的数据,消耗更多的进程。
答案 1 :(得分:1)
首先,您应该使用union all
而不是union
。所以从这个查询开始:
(SELECT receipt_ts, driver_id, equip_id, pos_lon, pos_lat, pos_timestamp
FROM log_messaging
WHERE equip_id = '207' AND tran_type = 'T.2.12.0'
ORDER BY receipt_ts DESC
LIMIT 1
) UNION ALL
(SELECT receipt_ts, driver_id, equip_id, pos_lon, pos_lat, pos_timestamp
FROM log_messaging
WHERE equip_id = '212' AND tran_type = 'T.2.12.0'
ORDER BY receipt_ts DESC
LIMIT 1
) UNION ALL
(SELECT receipt_ts, driver_id, equip_id, pos_lon, pos_lat, pos_timestamp
FROM log_messaging
WHERE equip_id = '213' AND tran_type = 'T.2.12.0'
ORDER BY receipt_ts DESC
LIMIT 1
) ;
这是一个合理的查询。我建议您在log_messaging(tran_type, equip_id, recipt_ts)
上创建一个索引。