Spring/JHipster QueryService如何构建深层关联Specification查询

问题根因&修复方案 #

1. 最直接的错误：Specification条件未生效 #

你在处理grandParentName过滤时，调用specification.and()后没有将生成的新Specification重新赋值给原变量，导致这部分过滤条件直接被丢弃，同时存在变量拼写错误、方法漏返回值的问题：

// 错误代码片段
if (criteria.getGrandParentName() != null) {
    // 此处仅生成条件，但没有赋值回原变量，条件直接丢失
    specification.and(buildJoinSpecification(criteria.getGrandParentName(), Child_.parent, Parent_.grandParent, GrandParent_.name));
}
// 变量拼写错误：speficiation → specification
speficiation = specification.and(buildReferringEntitySpecification(criteria.getParentName(), Child_.parent, Parent_.name));

修正后代码：

private Specification<Child> createSpecification(ChildCriteria criteria) {
    Specification<Child> specification = Specification.where(null);
    if (criteria != null) {
        if (criteria.getParentName() != null) {
            // 修正拼写错误
            specification = specification.and(buildReferringEntitySpecification(criteria.getParentName(), Child_.parent, Parent_.name));
        }
        if (criteria.getGrandParentName() != null) {
            // 新增赋值逻辑，把生成的过滤条件合并到原specification
            specification = specification.and(buildJoinSpecification(criteria.getGrandParentName(), Child_.parent, Parent_.grandParent, GrandParent_.name));
        }
    }
    // 补全返回值
    return specification;
}

2. buildJoinSpecification适配性不足 #

你当前自定义的buildJoinSpecification仅支持contains类型的过滤，没有覆盖equals/startsWith/endsWith等StringFilter的常用操作，会导致部分过滤请求不生效，同时指定左连接可以避免关联为空时的数据被意外过滤：
修正后的ExtendedQueryService代码：

import javax.persistence.criteria.JoinType;

public class ExtendedQueryService<ENTITY> extends QueryService<ENTITY>{
    protected <REFERENCE, JOIN, FILTER extends Comparable<? super FILTER>> Specification<ENTITY> buildJoinSpecification(StringFilter filter, SingularAttribute<? super ENTITY, REFERENCE> reference, SingularAttribute<REFERENCE, JOIN> joinField, SingularAttribute<JOIN, FILTER> valueField) {
        Specification<ENTITY> result = Specification.where((Specification) null);
        if (filter == null) {
            return result;
        }
        // 覆盖全量StringFilter操作逻辑
        if (filter.getEquals() != null) {
            result = result.and((root, query, builder) -> 
                builder.equal(root.join(reference, JoinType.LEFT).join(joinField, JoinType.LEFT).get(valueField), filter.getEquals())
            );
        }
        if (filter.getContains() != null) {
            result = result.and((root, query, builder) -> 
                builder.like(
                    builder.lower(root.join(reference, JoinType.LEFT).join(joinField, JoinType.LEFT).get(valueField)), 
                    "%" + filter.getContains().toLowerCase() + "%"
                )
            );
        }
        if (filter.getStartsWith() != null) {
            result = result.and((root, query, builder) -> 
                builder.like(
                    builder.lower(root.join(reference, JoinType.LEFT).join(joinField, JoinType.LEFT).get(valueField)), 
                    filter.getStartsWith().toLowerCase() + "%"
                )
            );
        }
        if (filter.getEndsWith() != null) {
            result = result.and((root, query, builder) -> 
                builder.like(
                    builder.lower(root.join(reference, JoinType.LEFT).join(joinField, JoinType.LEFT).get(valueField)), 
                    "%" + filter.getEndsWith().toLowerCase()
                )
            );
        }
        return result;
    }
}

3. 实体类注解错误修正 #

你的实体类中存在属性名大小写不匹配的问题，会导致关联查询失败：

• Child类的@QueryInit("GrandParent")修改为@QueryInit("grandParent")，对应Parent类的属性名是小写开头
• Parent类的@OneToMany(mappedBy = "Parent")修改为@OneToMany(mappedBy = "parent")，对应Child类的parent属性名
• GrandParent类的@OneToMany(mappedBy = "GrandParent")修改为@OneToMany(mappedBy = "grandParent")，对应Parent类的grandParent属性名

完成以上三处修改后，多层关联过滤逻辑即可正常运行。

内容的提问来源于stack exchange，提问作者derOtterDieb