BeetlSQL 3.37发布 新增多对多自动关联查询

BeetlSQL 自 2015 年起便投入研发，其核心模块及周边生态组件全部自主实现，未依赖任何第三方 ORM 内核。框架的设计目标极为明确：使开发者使用顺手、维护轻松、运行高效。简而言之，BeetlSQL 旨在提供一款学习成本极低、适用场景广泛且支持高度定制的数据库访问框架（Java ORM 框架）。

接下来，让我们一同了解本次版本更新带来的几项实用变化。

• 修复了 SQL Server 在特定场景下无法正确读取表元数据的缺陷
• Fetch 功能新增对多对多关系的自动加载支持
• 底层模板引擎 Beetl 升级至最新的 3.20.1 版本

Fetch 方面新增了 @FetchMany2Many 注解，并配套提供 FetchMany2ManyAction 类（需结合 @Builder 使用）。通过中间表自动拉取多对多关联数据，现在变得非常自然流畅。以下为注解的定义：

@Retention(RetentionPolicy.RUNTIME)
@Target(value = {ElementType.METHOD, ElementType.FIELD})
@Builder(FetchMany2ManyAction.class)
public @interface FetchMany2Many {
    Class tableClass();       // 中间表对应的实体类
    String fromAttr();        // 当前实体在中间表中的外键字段名
    String toAttr();          // 关联实体在中间表中的外键字段名
    String enableOn() default ""; // 启用条件表达式（可选）
}

在实际使用中，代码显得格外简洁：

@Table(name="sys_user")
public class MyUser {
    @AutoID
    Integer id;
    String name;
    @FetchMany2Many(tableClass = UserRole.class, fromAttr = "userId", toAttr = "roleId")
    List roles;
}

BeetlSQL 核心模块说明

BeetlSQL 官方扩展插件列表

启用加速插件后的性能对比显示：最高可达到原生 JDBC 性能的一半左右。下表列出了各类常见操作的具体数据。

接下来，展示更全面的 JMH 基准测试结果（单位：ops/ms），涵盖多个主流 Java ORM 框架的对比：

Benchmark                                  Mode  Cnt     Score      Error   Units
Beetl.addEntity                           thrpt    5   133.190 ±   38.548  ops/ms
Beetl.complexMapping                      thrpt    5   236.289 ±  288.370  ops/ms
Beetl.executeJdbcSql                      thrpt    5   298.657 ±   80.480  ops/ms
Beetl.executeTemplateSql                  thrpt    5   251.418 ±   77.757  ops/ms
Beetl.getAll                              thrpt    5    14.531 ±    0.605  ops/ms
Beetl.getEntity                           thrpt    5   359.275 ±   85.099  ops/ms
Beetl.lambdaQuery                         thrpt    5   187.228 ±   59.697  ops/ms
Beetl.one2Many                            thrpt    5   178.041 ±  181.829  ops/ms
Beetl.pageQuery                           thrpt    5   148.982 ±   39.327  ops/ms
Beetl.sqlFile                             thrpt    5   249.206 ±  100.083  ops/ms
DBVisitor.addEntity                       thrpt    5    99.294 ±   33.564  ops/ms
DBVisitor.executeJdbcSql                  thrpt    5   185.177 ±   50.993  ops/ms
DBVisitor.executeTemplateSql              thrpt    5     3.189 ±    0.748  ops/ms
DBVisitor.getAll                          thrpt    5     3.163 ±    0.980  ops/ms
DBVisitor.getEntity                       thrpt    5   125.683 ±   44.190  ops/ms
DBVisitor.lambdaQuery                     thrpt    5    94.314 ±   38.499  ops/ms
DBVisitor.pageQuery                       thrpt    5    49.954 ±   18.254  ops/ms
EasyQuery.addEntity                       thrpt    5    98.904 ±   39.890  ops/ms
EasyQuery.complexMapping                  thrpt    5    67.397 ±   59.919  ops/ms
EasyQuery.executeJdbcSql                  thrpt    5   243.160 ±   48.080  ops/ms
EasyQuery.getAll                          thrpt    5    14.896 ±    2.219  ops/ms
EasyQuery.getEntity                       thrpt    5   134.370 ±   67.962  ops/ms
EasyQuery.lambdaQuery                     thrpt    5   137.292 ±   53.438  ops/ms
EasyQuery.one2Many                        thrpt    5    75.083 ±   53.738  ops/ms
EasyQuery.pageQuery                       thrpt    5    84.596 ±   46.942  ops/ms
Flex.addEntity                            thrpt    5    86.484 ±   40.022  ops/ms
Flex.getAll                               thrpt    5     2.926 ±    0.993  ops/ms
Flex.getEntity                            thrpt    5    73.445 ±   39.437  ops/ms
Flex.pageQuery                            thrpt    5    33.999 ±   26.776  ops/ms
Jdbc.addEntity                            thrpt    5   225.223 ±  304.585  ops/ms
Jdbc.executeJdbcSql                       thrpt    5   691.616 ±  145.280  ops/ms
Jdbc.getAll                               thrpt    5    35.827 ±    7.295  ops/ms
Jdbc.getEntity                            thrpt    5   678.791 ±  165.527  ops/ms
Jpa.addEntity                             thrpt    5    53.219 ±   78.543  ops/ms
Jpa.executeJdbcSql                        thrpt    5    60.734 ±   61.299  ops/ms
Jpa.executeTemplateSql                    thrpt    5    64.478 ±   48.688  ops/ms
Jpa.getAll                                thrpt    5     4.710 ±    2.062  ops/ms
Jpa.getEntity                             thrpt    5    56.325 ±   53.744  ops/ms
Jpa.one2Many                              thrpt    5    92.168 ±  111.591  ops/ms
Jpa.pageQuery                             thrpt    5    58.756 ±   48.056  ops/ms
MyBatis.addEntity                         thrpt    5    40.533 ±   12.461  ops/ms
MyBatis.complexMapping                    thrpt    5   106.348 ±   57.688  ops/ms
MyBatis.executeTemplateSql                thrpt    5    46.953 ±   13.541  ops/ms
MyBatis.getAll                            thrpt    5     5.394 ±    0.502  ops/ms
MyBatis.getEntity                         thrpt    5    45.303 ±   11.928  ops/ms
MyBatis.lambdaQuery                       thrpt    5    23.756 ±   13.613  ops/ms
MyBatis.pageQuery                         thrpt    5    16.190 ±    5.682  ops/ms
MyBatis.sqlFile                           thrpt    5    42.558 ±   18.025  ops/ms
SqlUtils.addEntity                        thrpt    5   232.501 ±  332.656  ops/ms
SqlUtils.executeJdbcSql                   thrpt    5   682.610 ±  171.506  ops/ms
SqlUtils.getAll                           thrpt    5    44.353 ±    4.346  ops/ms
SqlUtils.getEntity                        thrpt    5   643.873 ±  131.587  ops/ms
Wood.addEntity                            thrpt    5   105.097 ±   79.961  ops/ms
Wood.executeJdbcSql                       thrpt    5   127.020 ±   48.142  ops/ms
Wood.executeTemplateSql                   thrpt    5   129.193 ±   44.455  ops/ms
Wood.getAll                               thrpt    5     2.051 ±    0.524  ops/ms
Wood.getEntity                            thrpt    5   126.104 ±   41.976  ops/ms
Wood.lambdaQuery                          thrpt    5   125.013 ±   43.991  ops/ms
Wood.pageQuery                            thrpt    5   220.101 ±   90.028  ops/ms
Wood.sqlFile                              thrpt    5   132.431 ±   50.141  ops/ms