Optimizing Batch Insert Performance with MyBatis-Plus: Configuration, Custom Methods, and Multithreading

Background

When synchronizing tens of thousands of base records from one system to another, using MyBatis-Plus's saveBatch() method without proper JDBC configuration can take many minutes per run.

The main cause is the missing rewriteBatchedStatements=true parameter in the MySQL connection URL.

Enabling Batch Rewrite

Add the following to the JDBC URL:

jdbc:mysql://
host
/
database
?useUnicode=true&characterEncoding=UTF8&allowMultiQueries=true&rewriteBatchedStatements=true

After enabling this flag, inserting 10,000 rows takes only a few hundred milliseconds.

Optimization Methods

2.1 MyBatis-Plus Built‑in saveBatch()

Entity definition:

@Data
@TableName("test_user")
public class TestUser implements Serializable {
    private String id;
    private String name;
    private String managerId;
    private String salary;
    private String age;
    private String departId;
    private String remark;
    private String province;
}

Mapper:

public interface TestUserMapper extends BaseMapper
{}

Service interface:

public interface ITestUserService extends IService
{}

Unit‑test inserting 200,000 rows with saveBatch() takes about 10 seconds.

2.2 Custom Batch Insert/Update

Create a generic RootMapper extending BaseMapper with custom methods:

public interface RootMapper
extends BaseMapper
{
    int insertBatch(@Param("list") Collection
batchList);
    int updateBatch(@Param("list") Collection
batchList);
}

Implement custom SQL injection classes:

public class InsertBatchColumn extends AbstractMethod {
    // builds INSERT ... VALUES script using MyBatis foreach
}

public class UpdateBatchColumn extends AbstractMethod {
    // builds UPDATE ... WHERE script using MyBatis foreach
}

Register them via a custom injector:

@Configuration
public class MybatiesPlusConfig {
    @Bean
    public MysqlInjector sqlInjector() {
        return new MysqlInjector();
    }
}

Now the mapper can call insertBatch() and updateBatch() directly.

2.3 Multithreaded Updates with saveBatch()

Using a fixed thread pool of 5 threads, each thread processes a batch of 5,000 records and calls userService.saveBatch(batchList) . Total time drops to ~3 seconds.

2.4 Multithreaded Custom Batch Insert

Similar thread pool, but each thread invokes testUserMapper.insertBatch(filteredList) . Execution time is about 2 seconds.

Summary

Setting rewriteBatchedStatements=true makes the built‑in batch insert fast enough for most cases (≈4 seconds for 200k rows). When further speed is needed, multithreading reduces the time to ~1 second, and custom batch methods provide comparable performance while offering more control.