优化 MySQL GROUP BY 查询

Question

我需要优化这个查询：

SELECT 
    cp.id_currency_purchase,
    MIN(cr.buy_rate),
    MAX(cr.buy_rate)
FROM currency_purchase cp
JOIN currency_rate cr ON cr.id_currency = cp.id_currency AND cr.id_currency_source = cp.id_currency_source
WHERE cr.rate_date >= cp.purchase_date
GROUP BY cp.id_currency_purchase

现在大约需要 0.5 秒，这对我来说太长了（它要么经常更新，所以查询缓存对我没有帮助）。

这是 EXPLAIN EXTENDED 输出：

+------+-------------+-------+-------+--------------------------------+-------------+---------+-------------------------+-------+----------+-------------+
| id   | select_type | table | type  | possible_keys                  | key         | key_len | ref                     | rows  | filtered | Extra       |
+------+-------------+-------+-------+--------------------------------+-------------+---------+-------------------------+-------+----------+-------------+
|    1 | SIMPLE      | cp    | index | id_currency_source,id_currency | PRIMARY     | 8       | NULL                    |     9 |   100.00 |             |
|    1 | SIMPLE      | cr    | ref   | id_currency,id_currency_source | id_currency | 8       | lokatnik.cp.id_currency | 21433 |   100.00 | Using where |
+------+-------------+-------+-------+--------------------------------+-------------+---------+-------------------------+-------+----------+-------------+

stackoverflow 上有一些关于一个表 GROUP BY 优化的好问题（例如：https://stackoverflow.com/a/11631480/1320081），但我真的不知道如何优化基于 JOIN 的表的查询。

如何加快速度？

Answer 1

此查询的最佳索引：

SELECT cp.id_currency_purchase, MIN(cr.buy_rate), MAX(cr.buy_rate)
FROM currency_purchase cp JOIN
     currency_rate cr
     ON cr.id_currency = cp.id_currency AND
        cr.id_currency_source = cp.id_currency_source
WHERE cr.rate_date >= cp.purchase_date
GROUP BY cp.id_currency_purchase;

是：currency_rate(id_currency, id_currency_source, rate_date, buy_rate)。你也可以试试currency_purchase(id_currency, id_currency_source, rate_date)。