I have a very simple table with two columns, but has 4.5M rows.
我有一个非常简单的表,有两列,但有4.5M行。
CREATE TABLE `content_link` (
`category_id` mediumint(8) unsigned NOT NULL,
`content_id` int(10) unsigned NOT NULL,
PRIMARY KEY (`content_id`,`category_id`),
KEY `content_id` (`content_id`,`category_id`)
) ENGINE=MyISAM;
When I run a simple query like:
当我运行一个简单的查询,如:
SELECT
*
FROM
content_link
WHERE
category_id = '11';
mysql spikes the CPU and takes 2-5 seconds before returning about 10 rows. The data is spread very evenly across the table and I'm accessing indexed fields (I've also analyzed/optimized the table and I never change the content of the table), so what reason is there for the query to take so long?
mysql激活CPU并在返回大约10行之前需要2-5秒。数据在表中非常均匀地传播,我正在访问索引字段(我还分析/优化了表,我从不更改表的内容),那么查询需要这么长时间的原因是什么?
Edit: It seems navicat was lying to me and my primary key was not actually keyed in the right order as it was displaying the table to me.
编辑:似乎navicat对我撒谎,我的主键实际上没有按正确的顺序键入,因为它正在向我显示表格。
3 个解决方案
#1
category_id
is not the first column in any index.
category_id不是任何索引中的第一列。
Recreate your secondary key as follows:
重新创建您的二级密钥如下:
UNIQUE KEY `ix_category_id` (`category_id`, `content_id`)
Note the column order, it matters.
注意列顺序,这很重要。
#2
The UNIQUE KEY ordering is a good solution and you should add a partition strategy to your table too.
UNIQUE KEY排序是一个很好的解决方案,您也应该在表中添加分区策略。
By partitioning the table in fragments, MySQL will query the specific partition with the right data set. I've applied and I had excellent results.
通过将表分割成片段,MySQL将使用正确的数据集查询特定分区。我申请了,我的成绩非常好。
CREATE TABLE ti (id INT, amount DECIMAL(7,2), tr_date DATE)
ENGINE=INNODB
PARTITION BY HASH( MONTH(tr_date) )
PARTITIONS 6;
You need MySQL 5.1.
你需要MySQL 5.1。
Try http://dev.mysql.com/doc/refman/5.1/en/partitioning.html
#3
You are not using the index. When you have a composite index like (content_id, category_id)
you can use the index by using content_id
or you can use content_id
and category_id
. You can't use category_id
and utilize the index.
您没有使用索引。如果您有像(content_id,category_id)这样的复合索引,则可以使用content_id来使用索引,也可以使用content_id和category_id。您不能使用category_id并使用索引。
Try changing:
KEY `content_id` (`content_id`, `category_id`)
to
KEY `category_id` (`category_id`, `content_id`)
#1
category_id
is not the first column in any index.
category_id不是任何索引中的第一列。
Recreate your secondary key as follows:
重新创建您的二级密钥如下:
UNIQUE KEY `ix_category_id` (`category_id`, `content_id`)
Note the column order, it matters.
注意列顺序,这很重要。
#2
The UNIQUE KEY ordering is a good solution and you should add a partition strategy to your table too.
UNIQUE KEY排序是一个很好的解决方案,您也应该在表中添加分区策略。
By partitioning the table in fragments, MySQL will query the specific partition with the right data set. I've applied and I had excellent results.
通过将表分割成片段,MySQL将使用正确的数据集查询特定分区。我申请了,我的成绩非常好。
CREATE TABLE ti (id INT, amount DECIMAL(7,2), tr_date DATE)
ENGINE=INNODB
PARTITION BY HASH( MONTH(tr_date) )
PARTITIONS 6;
You need MySQL 5.1.
你需要MySQL 5.1。
Try http://dev.mysql.com/doc/refman/5.1/en/partitioning.html
#3
You are not using the index. When you have a composite index like (content_id, category_id)
you can use the index by using content_id
or you can use content_id
and category_id
. You can't use category_id
and utilize the index.
您没有使用索引。如果您有像(content_id,category_id)这样的复合索引,则可以使用content_id来使用索引,也可以使用content_id和category_id。您不能使用category_id并使用索引。
Try changing:
KEY `content_id` (`content_id`, `category_id`)
to
KEY `category_id` (`category_id`, `content_id`)