SQL查询需要的时间太长 - 临时表

Question

我在运行MySQL的查询时遇到问题。感谢您的帮助，了解如何进行优化。

讨论中的查询表上运行约8000行，看起来像这样：

price_id int(5) Primary Key with auto increment 
product_id int(5) 
price_amount float 
price_date date 

基本上，它拥有的几款产品的历史价格，这样就可以生成相关的任何历史性的日子发票。

有问题的查询必须提供与特定日期相关的产品列表的最新价格。例如，获得的product_id 1,2和4，以9月1日的相关价格，这将是：

SELECT * 
FROM prices 
WHERE price_id IN (
    SELECT max(prices.price_id) 
    FROM prices 
    WHERE product_id in (1,2,4) AND price_date <= '2016-09-01' 
    GROUP BY product_id 
) 

现在，这个查询aaaaaages运行。它有所不同，但现在，例如，对于3个product_ids，运行大约需要58秒。通常情况下，发票有大约120〜产品在他们（需要几分钟的时间才能返回结果），所以你可以想像的麻烦..

我异形在phpMyAdmin查询，基本上它看起来像这样：

Status      Time 
Sorting result    0.000050 
Sending data     0.000025 
executing      0.000009 
Copying to tmp table   0.008586 
Sorting result    0.000052 
Sending data     0.000026 
executing      0.000006 
Copying to tmp table   0.008118 
Sorting result    0.000057 
Sending data     0.000029 
executing      0.000011 
Copying to tmp table   0.007498 
Sorting result    0.000047 
Sending data     0.000021 
executing      0.000005 
Copying to tmp table   0.008479 
Sorting result    0.000056 
Sending data     0.000031 
executing      0.000011 
Copying to tmp table   0.007371 
Sorting result    0.000059 
Sending data     0.000031 
executing      0.000011 
Copying to tmp table   0.006702 
Sorting result    0.000045 
Sending data     0.000019 
executing      0.000005 
Copying to tmp table   0.005319 
Sorting result    0.000034 
Sending data     0.000015 
executing      0.000005 
Copying to tmp table   0.005302 
Sorting result    0.000035 
Sending data     0.000016 
executing      0.000005 
Copying to tmp table   0.005207 
Sorting result    0.000031 
Sending data     0.000014 
executing      0.000005 
Copying to tmp table   0.005243 
Sorting result    0.000034 
Sending data     0.000016 
executing      0.000005 
Copying to tmp table   0.005236 
Sorting result    0.000035 
Sending data     0.000016 
executing      0.000005 
Copying to tmp table   0.005185 
Sorting result    0.000035 
Sending data     0.000015 
executing      0.000005 
Copying to tmp table   0.005256 
Sorting result    0.000033 
Sending data     0.000017 
executing      0.000005 
Copying to tmp table   0.005160 
Sorting result    0.000025 
Sending data     0.000014 
executing      0.000005 
Copying to tmp table   0.005149 
Sorting result    0.000024 
Sending data     0.000013 
executing      0.000005 
Copying to tmp table   0.005356 
Sorting result    0.000038 
Sending data     0.000016 
executing      0.000005 
Copying to tmp table   0.005221 
Sorting result    0.000034 
Sending data     0.000016 
executing      0.000005 
Copying to tmp table   0.005189 
Sorting result    0.000033 
Sending data     0.000015 
executing      0.000005 
Copying to tmp table   0.005370 
Sorting result    0.000038 
Sending data     0.000017 
executing      0.000005 
Copying to tmp table   0.005208 
Sorting result    0.000035 
Sending data     0.000017 
executing      0.000006 
Copying to tmp table   0.005209 
Sorting result    0.000036 
Sending data     0.000054 
end       0.000019 
removing tmp table   0.000021 
end       0.000008 
query end      0.000006 
closing tables    0.000020 
freeing items     0.000052 
Waiting for query cache lock 0.000006 
freeing items     0.000852 
Waiting for query cache lock 0.000012 
freeing items     0.000004 
storing result in query cache 0.000019 
logging slow query   0.000005 
logging slow query   0.000010 
cleaning up     0.000009 

Showing rows 0 - 2 (3 total, Query took 58.6074 sec) 

所以我想通过数字猜测，大部分时间都花在创建临时表格上。 （最大堆和临时表大小参数都是16M的默认值）。

有没有人有任何想法如何我可以加快这个查询或设计一个新的查询，但做得更好，但效率更高？

Answer 1

尝试使用从（一DINAMIC）和所得到的表

select * from prices as a 
    inner join (  
     SELECT max(prices.price_id) as max_price_id 
     FROM prices 
     WHERE product_id in (1,2,4) AND price_date <= '2016-09-01' 
     GROUP BY product_id) t on t.max_price_id = a.price_id 

Answer 2

使用相关查询内部联接，有时MySQL已经在子查询的问题

SELECT prices.* 
FROM prices 
WHERE EXISTS (
    SELECT max(p.price_id) 
    FROM prices p 
    WHERE p.product_id in (1,2,4) AND p.price_date <= '2016-09-01' 
    AND prices.price_id = p.price_id 
    GROUP BY p.product_id 
) 

Answer 3

，如果你是自由的添加索引，通过（1）product_id，（2）price_date（降序）索引表，并且您可以执行以下操作：

SELECT * 
FROM prices 
WHERE product_id in (1,2,4) AND price_date <= '2016-09-01' 
ORDER BY price_date DESC 
LIMIT 1