场景
最近在做数据分析项目,里面有这样一个业务:把匹配的数据打上标签,放到新的索引中。
数据量:累计亿级的数据
使用场景:可能会单次查询大量的数据,但不会设置复杂的条件,且这些数据不会被再次修改
原来使用的数据库:ElasticSearch
问题:上面也说了我这里打上标记后,这些数据几乎不会再修改了。ES 是一个全文检索引擎,更适用于进行大量文本检索的情况。这里与我上面的使用场景就不太匹配了。
技术选型的考虑:改用战斗民族开发的 ClickHouse,它适用于 OLAP 也就是数据分析的场景,当数据写入后,通过不同维度不断挖掘、分析,发现其中的商业价值。ClickHouse 适用于读远大于写的情况。
此外,相比ES,ClickHouse 占用的硬盘空间更小,也有利于降低运维成本。
下面是我在尝试接入 ClickHouse 时的一些实践,以及关于 ClickHouse数组类型转换问题的解决方案。
关于 ClickHouse 更详细的知识参考:https://zhuanlan.zhihu.com/p/98135840
示例代码已经上传到了 Git,目前更新第 28 节:https://github.com/laolunsi/spring-boot-examples/
Mybatis + ClickHouse
以前一直用 Mybatis 去操作 MySQL,其实 Mybatis 还可以操作 ClickHouse,这里用 Druid 进行连接管理。
maven 配置
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
|
< dependency >
< groupId >com.alibaba</ groupId >
< artifactId >druid</ artifactId >
< version >1.2.5</ version >
</ dependency >
< dependency >
< groupId >ru.yandex.clickhouse</ groupId >
< artifactId >clickhouse-jdbc</ artifactId >
< version >0.2.6</ version >
</ dependency >
< dependency >
< groupId >org.mybatis.spring.boot</ groupId >
< artifactId >mybatis-spring-boot-starter</ artifactId >
< version >2.1.3</ version >
</ dependency >
|
配置文件:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
|
spring:
datasource:
type : com.alibaba.druid.pool.DruidDataSource
# 注意这里是自定义的配置,通过 JdbcParamConfig 来加载配置到 Spring 中
# 然后由 DruidConfig 来配置数据源
click:
driverClassName: ru.yandex.clickhouse.ClickHouseDriver
url: jdbc:clickhouse: //127 .0.0.1:8123 /test # ip:port/database
userName: default
password: default # 按照自己连接的 clickhouse 数据库来
initialSize: 10
maxActive: 100
minIdle: 10
maxWait: 6000
validationQuery: SELECT 1
|
加载配置项的类:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
|
@Component
@ConfigurationProperties (prefix = "spring.datasource.click" )
public class JdbcParamConfig {
private String userName;
private String password;
private String driverClassName ;
private String url ;
private Integer initialSize ;
private Integer maxActive ;
private Integer minIdle ;
private Integer maxWait ;
private String validationQuery;
// ignore getters and setters
}
|
配置 Druid:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
|
@Configuration
@MapperScan (basePackages = {
"com.aegis.analysis.clickhousestorage.dao"
})
public class DruidConfig {
@Resource
private JdbcParamConfig jdbcParamConfig ;
@Bean (name = "clickDataSource" )
public DataSource dataSource() throws ClassNotFoundException {
Class classes = Class.forName( "com.alibaba.druid.pool.DruidDataSource" );
DruidDataSource dataSource = (DruidDataSource) DataSourceBuilder
.create()
.driverClassName(jdbcParamConfig.getDriverClassName())
.type(classes)
.url(jdbcParamConfig.getUrl())
.username(jdbcParamConfig.getUserName())
.password(jdbcParamConfig.getPassword())
.build();
dataSource.setMaxWait(jdbcParamConfig.getMaxWait());
dataSource.setValidationQuery(jdbcParamConfig.getValidationQuery());
return dataSource;
}
@Bean
public SqlSessionFactory clickHouseSqlSessionFactoryBean() throws Exception {
SqlSessionFactoryBean factory = new SqlSessionFactoryBean();
factory.setDataSource(dataSource());
// 实体 model的 路径 比如 com.order.model
factory.setTypeAliasesPackage( "com.example.clickhousedemo.model" );
//添加XML目录
ResourcePatternResolver resolver = new PathMatchingResourcePatternResolver();
factory.setMapperLocations(resolver.getResources( "classpath:mapper/*.xml" ));
//开启驼峰命名转换
factory.getObject().getConfiguration().setMapUnderscoreToCamelCase( true );
return factory.getObject();
}
}
|
定义一个 UserInfo 类,建表语句如下:
1
2
3
4
5
6
7
8
9
|
CREATE TABLE test. user (
`id` Int16,
` name ` String,
`score` Float32,
`score2` Float64,
`state` Int8,
`createTime` DateTime,
`ranks` Array(UInt8)
) ENGINE = MergeTree() ORDER BY id;
|
实体类:
1
2
3
4
5
6
7
8
9
10
11
12
13
|
public class UserInfo {
private Integer id; // int16
private String name; // String
private Float score; // float16
private Double score2; // float32
private Boolean state; // int8
private Date createTime; // datetime
private Integer[] ranks; // Array - Array 类型需要进行类型转换
// 具体转换方法与配置参考 ClickArrayToIntHandler 类与 UserMapper.xml 中关于查询和插入时 ranks 字段的配置
// ignore getters and setters
}
|
DAO 和 Mapper 文件就按照连接 MYSQL 时的写法一样。
这里有个需要注意的点,ClickHouse 有个 Array 类型,可以用来存数组,就像 ES 一样。问题是类型转换需要自己定义。网上一些资料仅列出了基本类型的场景,我自己实现了一个转换器,可以参考一下:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
|
/**
* Java Int 数组与 ClockHouse Array Int 转换器
* @version 1.0
* @since 2019/11/14 9:59
*/
public class ClickArrayToIntHandler extends BaseTypeHandler<Integer[]> {
@Override
public void setNonNullParameter(PreparedStatement preparedStatement, int i, Integer[] integers, JdbcType jdbcType) throws SQLException {
preparedStatement.setObject(i, integers);
}
@Override
public Integer[] getNullableResult(ResultSet resultSet, String s) throws SQLException {
Object obj = resultSet.getObject(s);
return parseClickHouseArrayToInt(obj);
}
@Override
public Integer[] getNullableResult(ResultSet resultSet, int i) throws SQLException {
Object obj = resultSet.getObject(i);
return parseClickHouseArrayToInt(obj);
}
@Override
public Integer[] getNullableResult(CallableStatement callableStatement, int i) throws SQLException {
Object obj = callableStatement.getObject(i);
return parseClickHouseArrayToInt(obj);
}
private Integer[] parseClickHouseArrayToInt(Object obj) {
if (obj instanceof ClickHouseArray) {
int [] res = new int [ 0 ];
try {
res = ( int []) ((ClickHouseArray) obj).getArray();
} catch (SQLException ex) {
ex.printStackTrace();
}
if (res != null && res.length > 0 ) {
Integer[] resI = new Integer[res.length];
for ( int i = 0 ; i < res.length; i++) {
resI[i] = res[i];
}
return resI;
}
}
return new Integer[ 0 ];
}
}
|
DAO.xml 也给一个示例:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
|
<? xml version = "1.0" encoding = "UTF-8" ?>
<!DOCTYPE mapper PUBLIC "-//mybatis.org//DTD Mapper 3.0//EN" "http://mybatis.org/dtd/mybatis-3-mapper.dtd">
< mapper namespace = "com.aegis.analysis.clickhousestorage.dao.UserInfoMapper" >
< resultMap id = "BaseResultMap" type = "com.example.clickhousedemo.model.UserInfo" >
< id column = "id" property = "id" />
< result column = "name" property = "name" />
< result column = "name" property = "name" />
< result column = "score" property = "score" />
< result column = "score2" property = "score2" />
< result column = "state" property = "state" />
< result column = "createTime" property = "createTime" />
<!-- <result column="ranks" property="ranks" jdbcType="JAVA_OBJECT" javaType="java.lang.Object" />-->
< result column = "ranks" property = "ranks" typeHandler = "com.example.clickhousedemo.dao.ClickArrayToIntHandler" />
</ resultMap >
< sql id = "Base_Column_List" >
*
</ sql >
< insert id = "saveData" parameterType = "com.aegis.analysis.clickhousestorage.model.UserInfo" >
INSERT INTO user
(id,name, score, score2, state, createTime, ranks)
VALUES
(#{id},#{name}, #{score}, #{score2}, #{state}, #{createTime}, #{ranks, jdbcType=ARRAY,
typeHandler=com.example.clickhousedemo.dao.ClickArrayToIntHandler})
</ insert >
< select id = "selectById" resultMap = "BaseResultMap" >
select
< include refid = "Base_Column_List" />
from user
where id = #{id}
limit 1
</ select >
< select id = "selectList" resultMap = "BaseResultMap" >
select
< include refid = "Base_Column_List" />
from user
</ select >
</ mapper >
|
具体代码可以去我的 Git 仓库里查看,还有 SpringBoot 整合其他中间件技术的示例,欢迎 Star!
https://github.com/laolunsi/spring-boot-examples
以上就是如何将Mybatis连接到ClickHouse的详细内容,更多关于Mybatis连接到ClickHouse的资料请关注服务器之家其它相关文章!
原文链接:http://www.eknown.cn/index.php/spring-boot/clickhouse.html