我正在尝试根据购买的商品数量“捆绑”所有客户,并显示每个箱柜的数量.我试图看看有多少人(account_id)购买了一件商品,购买了两件商品,一直购买了九件商品,然后购买了十件商品.
这是我正在使用的查询 – 为了它的价值,我希望查询对销售进行全表扫描以生成结果,但整个过程需要永远!
我来自Oracle背景,我像在Oracle中一样编写查询.
select thecnt
, count(*)
from (select count(*)
, case when count(*) >= 10 then 'tenormore' else cast(count(*) as char) end thecnt
from sales
where created >= SUBDATE( CURRENT_DATE(), INTERVAL 60 DAY )
group by account_id) sub
group by thecnt
order by thecnt;
在处理子查询时,mysql中有任何陷阱吗?
解释计划
+----+-------------+-------------------+-------+---------------+---------+---------+------+---------+----------+-----------------------------------------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+----+-------------+-------------------+-------+---------------+---------+---------+------+---------+----------+-----------------------------------------------------------+
| 1 | PRIMARY | <derived2> | ALL | NULL | NULL | NULL | NULL | 2143248 | 100.00 | Using temporary; Using filesort |
| 2 | DERIVED | sales | range | created | created | 8 | NULL | 2012492 | 100.00 | Using where; Using index; Using temporary; Using filesort |
+----+-------------+-------------------+-------+---------------+---------+---------+------+---------+----------+-----------------------------------------------------------+
2 rows in set, 1 warning (1 hour 4 min 6.14 sec)
mysql> describe sales;
+-----------------+---------------------+------+-----+---------+-------+
| Field | Type | Null | Key | Default | Extra |
+-----------------+---------------------+------+-----+---------+-------+
| account_id | char(36) | NO | PRI | NULL | |
| created | datetime | NO | MUL | NULL | |
| histogram_value | bigint(20) unsigned | NO | PRI | NULL | |
+-----------------+---------------------+------+-----+---------+-------+
最佳答案 我没有看到您的查询有任何特别的错误.查询速度慢的原因是因为它需要使用临时表和filesort.严重加速此查询的唯一方法是修改MySQL设置以分配更多内存,以避免将磁盘用于这些进程.
Here’s a spot on article covering the pertinent settings.
编辑:执行此操作后,您还可以通过指定要计数的精确列而不是COUNT(*)以及其他一些小调整来节省内存,正如其他一些人所提到的那样.您希望获得尽可能小的数据集以充分利用您的记忆.但我认为除非你分配更多内存,否则整个问题不会消失.