SELECT
t1.theater_id,
sum(t2.full_sale_price * t1.full_tickets +
t2.half_sale_price * t1.half_tickets) as Gross,
sum(t2.full_cost_price * t1.full_tickets +
t2.half_cost_price * t1.half_tickets) as Cost,
Round(sum((t2.full_sale_price * t1.full_tickets +
t2.half_sale_price * t1.half_tickets) * t3.rate) / 100,
2) as IPG,
(sum(t2.full_sale_price * t1.full_tickets +
t2.half_sale_price * t1.half_tickets)
- Round(sum((t2.full_sale_price * t1.full_tickets +
t2.half_sale_price * t1.half_tickets) * (CASE
WHEN ISNULL(t3.rate) THEN 0
ELSE t3.rate
END)) / 100,
2)) as NetRev,
CASE
WHEN (t1.method = '2') THEN sum(0.00)
ELSE 0
END as sms,
sum((CASE
WHEN (t4.type = '1') THEN ((t1.full_tickets * 2) + (t1.half_tickets))
ELSE (t1.full_tickets + t1.half_tickets)
END) * 4.00) as maintenance,
sum((((t2.full_sale_price * t1.full_tickets +
t2.half_sale_price * t1.half_tickets)
- (Round((t2.full_sale_price * t1.full_tickets +
t2.half_sale_price * t1.half_tickets) * (CASE
WHEN ISNULL(t3.rate) THEN 0
ELSE t3.rate
END) / 100,
2))) - (t2.full_cost_price * t1.full_tickets +
t2.half_cost_price * t1.half_tickets))
- (CASE
WHEN (t1.method = '2') THEN 0.00
ELSE 0
END)
- (CASE
WHEN (t4.type = '1') THEN ((t1.full_tickets * 2) +
(t1.half_tickets))
ELSE (t1.full_tickets + t1.half_tickets)
END) * 4.00) as admincost,
sum(CASE
WHEN (t4.type = '1') THEN ((t1.full_tickets * 2) + (t1.half_tickets))
ELSE (t1.full_tickets + t1.half_tickets)
END) as totTk
FROM
`ratecard_rates` as t2 INNER JOIN
`reservation` as t1 ON t1.movie_id = t2.movie_id
AND t1.theater_id = t2.theater_id
AND t1.showtime_id = t2.showtime_id
AND t1.category_id = t2.category_id
AND t1.bx_date = t2.date_apply
LEFT JOIN
`paymentgateway` as t3 ON t1.paymentgateway_id = t3.id
LEFT JOIN
`theatercategories` as t4 ON t1.category_id = t4.id
WHERE
t1.status = '1' AND t1.method = '1'
AND t1.reservation_type = '1'
AND DATE(`tx_date`) BETWEEN DATE('2012-08-27') AND DATE('2012-08-27')
GROUP BY t1.theater_id
这是EXPLAIN
答案 0 :(得分:2)
"Using temporary; using filesort"
通常意味着在查询计划的这一点上,MySsql决定将临时集写入磁盘,遗憾的是,它并不总是能够使用您可能放在相关列上的索引。
根据我的经验,解决这个问题的唯一方法是尝试减少中期结果的设置大小。尝试并隔离发生这种情况的点:例如:用SELECT
替换整个COUNT(*)
以消除CASE位等,看看是什么产生了MySql想要“缓存”的2779行在磁盘上。
有关详细信息,请参阅this answer。
答案 1 :(得分:0)
您是盲目索引还是先检查了执行计划? CASE..END
子句有很多帮助但有时可能会破坏很多执行计划。检查this link它可以帮助您理解我在说什么。一种选择是创建存储过程并在较小的查询上拆分此查询。有时子查询可能会有所帮助,这是尝试的另一种选择。