sql如下
SELECT DISTINCT o.orders_id, o.oa_order_id,os.orders_status_name, o.order_type,
o.date_purchased AS add_date,dop.resource, dop.country_code
FROM dm_order_products AS dop
LEFT JOIN orders AS o ON o.orders_id=dop.orders_id
LEFT JOIN orders_total AS ot ON ot.orders_id=o.orders_id AND ot.class='ot_total'
LEFT JOIN orders_status AS os ON os.orders_status_id=o.orders_status
WHERE o.date_purchased >= '2014-01-31 10:00:00' AND o.date_purchased <= '2014-02-24 09:59:59'
ORDER BY o.orders_id DESC LIMIT 0, 20;因为需要在大结果集中order by 去重，再显示20条.
表特性是orders(o)表对dm_order_products(dop)表为一对多关系,而取出来的dop.country_code为一个订单号对应唯一值，由于表结构设计问题，每次查询该country_code都需要去dop查询。所以，每次查询都放大结果集，然后再去重，得到所要的结果集合。
explain
+----+-------------+-------+-------+----------------------------------+----------------------------+---------+-------------------------------+-------+----------------------------------------------+ | id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra | +----+-------------+-------+-------+----------------------------------+----------------------------+---------+-------------------------------+-------+----------------------------------------------+ | 1 | SIMPLE | o | range | PRIMARY,date_purchased | date_purchased | 9 | NULL | 952922 | Using where; Using temporary; Using filesort | | 1 | SIMPLE | ot | ref | idx_orders_total_orders_id,class | idx_orders_total_orders_id | 4 | banggood_work.o.orders_id | 3 | | | 1 | SIMPLE | os | ref | PRIMARY | PRIMARY | 4 | banggood_work.o.orders_status | 1 | | | 1 | SIMPLE | dop | ref | orders_id | orders_id | 4 | banggood_work.o.orders_id | 2 | | +----+-------------+-------+-------+----------------------------------+----------------------------+---------+-------------------------------+-------+----------------------------------------------+
索引情况使用正常，但是发现需要扫描一个大结果集.
profiling，执行时间为将近20s
mysql> show profile cpu,block io for query 1; +--------------------------------+-----------+----------+------------+--------------+---------------+ | Status | Duration | CPU_user | CPU_system | Block_ops_in | Block_ops_out | +--------------------------------+-----------+----------+------------+--------------+---------------+ | starting | 0.000025 | 0.000000 | 0.000000 | 0 | 0 | | Waiting for query cache lock | 0.000004 | 0.000000 | 0.000000 | 0 | 0 | | checking query cache for query | 0.000080 | 0.000000 | 0.000000 | 0 | 0 | | checking permissions | 0.000005 | 0.000000 | 0.000000 | 0 | 0 | | checking permissions | 0.000003 | 0.000000 | 0.000000 | 0 | 0 | | checking permissions | 0.000003 | 0.000000 | 0.000000 | 0 | 0 | | checking permissions | 0.000006 | 0.000000 | 0.000000 | 0 | 0 | | Opening tables | 0.000034 | 0.000000 | 0.000000 | 0 | 0 | | System lock | 0.000012 | 0.000000 | 0.000000 | 0 | 0 | | Waiting for query cache lock | 0.000024 | 0.000000 | 0.000000 | 0 | 0 | | init | 0.000046 | 0.000000 | 0.000000 | 0 | 0 | | optimizing | 0.000018 | 0.000000 | 0.000000 | 0 | 0 | | statistics | 0.000193 | 0.000000 | 0.000000 | 0 | 0 | | preparing | 0.000054 | 0.000000 | 0.000000 | 0 | 0 | | Creating tmp table | 0.000031 | 0.000000 | 0.000000 | 0 | 0 | | executing | 0.000004 | 0.000000 | 0.000000 | 0 | 0 | | Copying to tmp table | 12.491533 | 3.039538 | 3.107527 | 11896 | 824 | | Sorting result | 0.030709 | 0.034995 | 0.004000 | 16 | 496 | | Sending data | 0.000048 | 0.000000 | 0.000000 | 0 | 0 | | end | 0.000004 | 0.000000 | 0.000000 | 0 | 0 | | removing tmp table | 0.010108 | 0.000000 | 0.010998 | 8 | 32 | | end | 0.000013 | 0.000000 | 0.000000 | 0 | 0 | | query end | 0.000004 | 0.000000 | 0.000000 | 0 | 0 | | closing tables | 0.000012 | 0.000000 | 0.000000 | 0 | 0 | | freeing items | 0.000338 | 0.000000 | 0.000000 | 0 | 0 | | logging slow query | 0.000006 | 0.000000 | 0.000000 | 0 | 0 | | logging slow query | 0.000033 | 0.000000 | 0.000000 | 0 | 8 | | cleaning up | 0.000006 | 0.000000 | 0.000000 | 0 | 0 |
可以看到Copying to tmp table 占了大部分的cpu时间和io,最后sorting result占比重不大。
我们可以上面描述的结合特性，是否能够去掉Copying to tmp table 选项！因为是根据orders_id排序，取出最新的20条数据，如果我们在orders表中先把20条数据取出来，再和对应的表连接，这样一来，就将整个大结果Copying to tmp table 再排序这一步去掉！
看sql语句如下
SELECT DISTINCT o.orders_id, o.oa_order_id,os.orders_status_name, o.order_type,
o.date_purchased AS add_date,dop.resource, dop.country_code
FROM
(
SELECT *
FROM orders AS o
WHERE o.date_purchased >= '2014-01-31 10:00:00' AND o.date_purchased <= '2014-02-24 09:59:59'
ORDER BY o.orders_id DESC LIMIT 0, 20
) o
LEFT JOIN dm_order_products AS dop ON o.orders_id=dop.orders_id
LEFT JOIN orders_total AS ot ON ot.orders_id=o.orders_id AND ot.class='ot_total'
LEFT JOIN orders_status AS os ON os.orders_status_id=o.orders_status
ORDER BY o.orders_id DESC LIMIT 0, 20;
+----+-------------+------------+-------+----------------------------------+----------------------------+---------+-----------------+------+---------------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+------------+-------+----------------------------------+----------------------------+---------+-----------------+------+---------------------------------+
| 1 | PRIMARY | <derived2> | ALL | NULL | NULL | NULL | NULL | 20 | Using temporary; Using filesort |
| 1 | PRIMARY | dop | ref | orders_id | orders_id | 4 | o.orders_id | 2 | |
| 1 | PRIMARY | ot | ref | idx_orders_total_orders_id,class | idx_orders_total_orders_id | 4 | o.orders_id | 3 | |
| 1 | PRIMARY | os | ref | PRIMARY | PRIMARY | 4 | o.orders_status | 1 | |
| 2 | DERIVED | o | index | date_purchased | PRIMARY | 4 | NULL | 330 | Using where |
+----+-------------+------------+-------+----------------------------------+----------------------------+---------+-----------------+------+---------------------------------+可以发现，结果集瞬间缩小N倍,再查看profiling，执行时间变成0.01s
+--------------------------------+----------+----------+------------+--------------+---------------+ | Status | Duration | CPU_user | CPU_system | Block_ops_in | Block_ops_out | +--------------------------------+----------+----------+------------+--------------+---------------+ | starting | 0.000028 | 0.000000 | 0.000000 | 0 | 0 | | Waiting for query cache lock | 0.000008 | 0.000000 | 0.000000 | 0 | 0 | | checking query cache for query | 0.000148 | 0.000000 | 0.000000 | 0 | 0 | | checking permissions | 0.000010 | 0.000000 | 0.000000 | 0 | 0 | | checking permissions | 0.000006 | 0.000000 | 0.000000 | 0 | 0 | | checking permissions | 0.000006 | 0.000000 | 0.000000 | 0 | 0 | | checking permissions | 0.000010 | 0.000000 | 0.000000 | 0 | 0 | | Opening tables | 0.000052 | 0.000000 | 0.000000 | 0 | 0 | | System lock | 0.001299 | 0.000000 | 0.001000 | 0 | 8 | | optimizing | 0.000021 | 0.000000 | 0.000000 | 0 | 0 | | statistics | 0.000127 | 0.000000 | 0.000000 | 0 | 0 | | preparing | 0.000127 | 0.001000 | 0.000000 | 0 | 0 | | executing | 0.000009 | 0.000000 | 0.000000 | 0 | 0 | | Sorting result | 0.000012 | 0.000000 | 0.000000 | 0 | 0 | | Sending data | 0.002182 | 0.001000 | 0.001000 | 0 | 784 | | Waiting for query cache lock | 0.000010 | 0.000000 | 0.000000 | 0 | 0 | | Sending data | 0.000053 | 0.000000 | 0.000000 | 0 | 0 | | init | 0.000048 | 0.000000 | 0.000000 | 0 | 0 | | optimizing | 0.000010 | 0.000000 | 0.000000 | 0 | 0 | | statistics | 0.000315 | 0.000000 | 0.000000 | 0 | 0 | | preparing | 0.000025 | 0.000000 | 0.000000 | 0 | 0 | | Creating tmp table | 0.000043 | 0.000000 | 0.000000 | 0 | 0 | | executing | 0.000008 | 0.000000 | 0.000000 | 0 | 0 | | Copying to tmp table | 0.003390 | 0.002000 | 0.002000 | 0 | 0 | | Sorting result | 0.000028 | 0.000000 | 0.000000 | 0 | 0 | | Sending data | 0.000058 | 0.000000 | 0.000000 | 0 | 0 | | end | 0.000008 | 0.000000 | 0.000000 | 0 | 0 | | removing tmp table | 0.000013 | 0.000000 | 0.000000 | 0 | 0 | | end | 0.000008 | 0.000000 | 0.000000 | 0 | 0 | | query end | 0.000008 | 0.000000 | 0.000000 | 0 | 0 | | closing tables | 0.000007 | 0.000000 | 0.000000 | 0 | 0 | | removing tmp table | 0.000591 | 0.000000 | 0.000000 | 0 | 0 | | closing tables | 0.000021 | 0.000000 | 0.000000 | 0 | 0 | | freeing items | 0.000024 | 0.000000 | 0.000000 | 0 | 0 | | Waiting for query cache lock | 0.000007 | 0.000000 | 0.000000 | 0 | 0 | | freeing items | 0.000630 | 0.001000 | 0.000000 | 0 | 0 | | Waiting for query cache lock | 0.000010 | 0.000000 | 0.000000 | 0 | 0 | | freeing items | 0.000006 | 0.000000 | 0.000000 | 0 | 0 | | storing result in query cache | 0.000010 | 0.000000 | 0.000000 | 0 | 0 | | logging slow query | 0.000006 | 0.000000 | 0.000000 | 0 | 0 | | logging slow query | 0.000056 | 0.000000 | 0.000000 | 0 | 8 | | cleaning up | 0.000011 | 0.000000 | 0.000000 | 0 | 0 | +--------------------------------+----------+----------+------------+--------------+---------------+
可以看到Copying to tmp table已经被优化！执行时间提高了几百倍！
注意：
1. 如果是limit m,n这种语句类型，第一步要想到，是不是可以抽取一部分使用临时表优化后，再和外面的其他表进行优化
2. distinct和orders by 是天敌。在大结果集的sql语句中，尽量避免直接使用distinct.否则，mysql会将整个结果集copy导临时表，重新排序后，再返回！
