Optimizing Batch Insert Performance in MyBatis‑Plus with a Custom SQL Injector

The author describes a high‑frequency IoT data ingestion scenario where MySQL tables receive millions of rows, causing Kafka back‑pressure and long batch insert times (10‑20 seconds per 1500 rows). The default saveBatch() method only loops inserts one by one, even with rewriteBatchedStatements=true enabled.

To achieve true batch insertion, the article introduces MyBatis‑Plus's SQL injector mechanism. It shows the built‑in injectable methods and highlights InsertBatchSomeColumn as the method that generates a single INSERT statement for multiple rows.

Custom injector implementation:

public class MySqlInjector extends DefaultSqlInjector {
    @Override
    public List
getMethodList(Class
mapperClass) {
        List
methodList = super.getMethodList(mapperClass);
        // Add real batch insert method, ignoring fields filled only on UPDATE
        methodList.add(new InsertBatchSomeColumn(i -> i.getFieldFill() != FieldFill.UPDATE));
        return methodList;
    }
}

The injector is registered in Spring configuration:

@Configuration
public class MybatisPlusConfig {
    @Bean
    public MySqlInjector sqlInjector() {
        return new MySqlInjector();
    }
}

A mapper interface extending a custom CommonMapper gains the new method:

public interface CommonMapper
extends BaseMapper
{
    int insertBatchSomeColumn(List
entityList);
}

Usage example:

@Test
void contextLoads() {
    for (int i = 0; i < 5; i++) {
        User user = new User();
        user.setAge(10);
        user.setUsername("zhmsky");
        user.setEmail("[email protected]");
        userList.add(user);
    }
    long start = System.currentTimeMillis();
    userMapper.insertBatchSomeColumn(userList);
    long end = System.currentTimeMillis();
    System.out.println("Time:" + (end - start));
    userList.clear();
}

Performance tests compare the fake batch insert (original saveBatch ) with the real batch insert ( insertBatchSomeColumn ) on 50,000 records. The custom method reduces execution time by about 3 seconds and, when inserting 1,500 rows per batch, completes in roughly 650 ms.

Finally, the author adds a note encouraging readers to like, share, and follow the public account for more technical content.