今天帮开发人员优化SQL,发现一个以前不怎么在意的问题
SELECT
bond.store_no
FROM
bill_order_num_dtl bond
WHERE EXISTS
( select 1
from meetorder_staff_dtl b ,
meetorder_staff a
where b.operate_area = bond.store_no
and a.seq_no = b.seq_no
and a.user_code = 'lym'
) ;
查询需要2S,bill_order_num_dtl 数据量大概2W多条,meetorder_staff_dtl 50条左右,将EXISTS换为INNER JOIN后只需0.02S,差了100倍,EXISTS在命中率高的情况下查询速度较快,像这种需要判断的表的数据量较大,而条件表的数据量较少时不宜使用。INNER JOIN相对较稳定一些,不会随命中率的变化而影响性能。改成下面后,只需0.02S
SELECT bond.store_no
FROM bill_order_num_dtl bond
INNER JOIN meetorder_staff_dtl b on b.operate_area = bond.store_no
INNER JOIN meetorder_staff a on a.seq_no = b.seq_no
WHERE
a.user_code = 'lym'
转载于:https://blog.51cto.com/cyang/1562264