A read cache mechanism in mybatis

Caching feature for the query (if not heard UPDATE, INSERT statement what to cache are performed directly)

By default, mybatis cache enabled.

If you use the same session object called the same SELECT statement will return results directly from the cache, rather than re-query the database again.

Note: session or close call commit method after this session of the cache will be cleared

For example: The log output can be seen, the following code will only send a query sql

SqlSession sqlSession = MyBatisSqlSessionFactory.openSession();
OneToManyMapper mapper = sqlSession.getMapper(OneToManyMapper.class);
Tutor findTutorById1 = mapper.findTutorById(1);
System.out.println(findTutorById1);
Tutor findTutorById2 = mapper.findTutorById(1);
System.out.println(findTutorById2);

Log file output:

2019-10-16 19:15:39,870 [DEBUG] com.briup.mappers.OneToManyMapper.findTutorById - ==> Preparing: select * from tutors where tutor_id=?
2019-10-16 19:15:39,999 [DEBUG] com.briup.mappers.OneToManyMapper.findTutorById - ==> Parameters: 1(Integer)
2019-10-16 19:15:40,218 [DEBUG] com.briup.mappers.OneToManyMapper.findAddressById - ====> Preparing: select * from addresses where addr_id = ?
2019-10-16 19:15:40,218 [DEBUG] com.briup.mappers.OneToManyMapper.findAddressById - ====> Parameters: 1(Integer)
2019-10-16 19:15:40,220 [DEBUG] com.briup.mappers.OneToManyMapper.findAddressById - <==== Total: 1
2019-10-16 19:15:40,221 [DEBUG] com.briup.mappers.OneToManyMapper.findCoursesByTutor - ====> Preparing: select * from courses where tutor_id=?
2019-10-16 19:15:40,221 [DEBUG] com.briup.mappers.OneToManyMapper.findCoursesByTutor - ====> Parameters: 1(Integer)
2019-10-16 19:15:40,226 [DEBUG] com.briup.mappers.OneToManyMapper.findCoursesByTutor - <==== Total: 2
2019-10-16 19:15:40,226 [DEBUG] com.briup.mappers.OneToManyMapper.findTutorById - <== Total: 1

Tutor [tutorId=1, name=null, [email protected], phone=PhoneNumber [countryCode=123, stateCode=456, number=7890], address=Address [addrId=1, street=redSt, city=kunshan, state=W, zip=12345, country=china], courses=[Course [courseId=1, name=JavaSE, description=JavaSE, startDate=Thu Jan 10 00:00:00 CST 2019, endDate=Sun Feb 10 00:00:00 CST 2019], Course [courseId=3, name=MyBatis, description=MyBatis, startDate=Thu Jan 10 00:00:00 CST 2019, endDate=Wed Feb 20 00:00:00 CST 2019]]]

From the above logs can be analyzed, the select statement only made once, the second query using a first query cache

Secondary cache: between different caches may share session object data
1.mybatis-config.xml file to ensure <setting name = "cacheEnabled" value = "true" /> is set cache function is turned on, the default is open to true
2. mapping xml documents in the secondary cache, the cache function manual opening, a label is added to the root element: <cache />
3. after completion of a session query data, call the commit or close methods, the data will be entered into the second level cache, then the other session can share the cache data to the

Note: by default, the secondary cache is saved objects need to implement the serialization interface.

For example:
MyBatis-the config.xml:

<settings>
<setting name="cacheEnabled" value="true"/>
</settings>

xml mapping file:

<mapper namespace="com.briup.mappers.SpecialMapper">
<cache/>
<select> ..</select>
<select> ..</select>
<select> ..</select>
</mapper>

Test code:

@Test
public void test_cache2(){
SqlSession session1 = null;
SqlSession session2 = null;
try {
session1 = MyBatisSqlSessionFactory.openSession();
session2 = MyBatisSqlSessionFactory.openSession();

SpecialMapper mapper1 = session1.getMapper(SpecialMapper.class);
SpecialMapper mapper2 = session2.getMapper(SpecialMapper.class);

User user1 = mapper1.findUserById(56);
System.out.println(user1);
session1.commit();

User user2 = mapper2.findUserById(56);
System.out.println(user2);
session2.commit();
} catch (Exception e) {
e.printStackTrace();
}finally {
if(session1!=null)session1.close();
if(session2!=null)session2.close();
}
}

Log file output:

2019-10-16 19:29:22,443 [DEBUG] com.briup.mappers.OneToManyMapper - Cache Hit Ratio [com.briup.mappers.OneToManyMapper]: 0.0
2019-10-16 19:29:23,449 [DEBUG] com.briup.mappers.OneToManyMapper.findTutorById - ==> Preparing: select * from tutors where tutor_id=?
2019-10-16 19:29:23,648 [DEBUG] com.briup.mappers.OneToManyMapper.findTutorById - ==> Parameters: 1(Integer)
2019-10-16 19:29:23,805 [DEBUG] com.briup.mappers.OneToManyMapper - Cache Hit Ratio [com.briup.mappers.OneToManyMapper]: 0.0
2019-10-16 19:29:23,805 [DEBUG] com.briup.mappers.OneToManyMapper.findAddressById - ====> Preparing: select * from addresses where addr_id = ?
2019-10-16 19:29:23,806 [DEBUG] com.briup.mappers.OneToManyMapper.findAddressById - ====> Parameters: 1(Integer)
2019-10-16 19:29:23,808 [DEBUG] com.briup.mappers.OneToManyMapper.findAddressById - <==== Total: 1
2019-10-16 19:29:23,808 [DEBUG] com.briup.mappers.OneToManyMapper - Cache Hit Ratio [com.briup.mappers.OneToManyMapper]: 0.0
2019-10-16 19:29:23,809 [DEBUG] com.briup.mappers.OneToManyMapper.findCoursesByTutor - ====> Preparing: select * from courses where tutor_id=?
2019-10-16 19:29:23,809 [DEBUG] com.briup.mappers.OneToManyMapper.findCoursesByTutor - ====> Parameters: 1(Integer)
2019-10-16 19:29:23,814 [DEBUG] com.briup.mappers.OneToManyMapper.findCoursesByTutor - <==== Total: 2
2019-10-16 19:29:23,814 [DEBUG] com.briup.mappers.OneToManyMapper.findTutorById - <== Total: 1

2019-10-16 19:29:23,825 [DEBUG] com.briup.mappers.OneToManyMapper - Cache Hit Ratio [com.briup.mappers.OneToManyMapper]: 0.25

As can be seen, when the object sqlSession2 second query, cache hit

2019-10-16 19: 29: 23,825 [DEBUG] com.briup.mappers.OneToManyMapper - Cache Hit Ratio [com.briup.mappers.OneToManyMapper]: 0.25 (cache hit ratio 0.25)

Secondary cache added that
　　all select statements 1. mapped statement file will be cached
　　all insert, update and delete statements 2. mapped statement file will flush the cache
　　3. The cache will use a Least Recently Used (LRU, Least Recently Used a) algorithm for eviction.
　　4. The cache is refreshed according to a specified time interval.
　　The cache stores objects 1024

cache tag common attributes:
<cache
eviction = "the FIFO" <- recovery strategy FIFO ->!
flushInterval = "60000" <- Auto refresh time 60s ->!
size = "512" <-! referenced object buffer up to 512 ->
readOnly = "to true" /> <-! objects can not be written to true indicates out, i.e., may not be serialized representation can write out to false, i.e., can be serialized, the default value is false ->

A read cache mechanism in mybatis

Guess you like