Spring Data -Specification用法和常用查询方法（in，join，equal等）

前言
入门例子
Repository层常用写法
Specification 的用法
总结

前言

在这一年里技术更新，从使用Mybatis转为Spring Data,总体感受是终于不用在自己写映射了，也可以少写方法和对应字段了。接下来总结在工作中常用的查询方式和方法，例如equal，join,in等。积少成多，在这里只收藏了用过查询与方法（如果方法可以用得更有，有更多的用法和不足之处请联系我）。

入门例子

controller层，这里是个简单查询获取所有用户并分页

  @GetMapping("/list")
    @ApiOperation(value = "所有用户列表")
    public Result list(@RequestParam Map<String, Object> params) {
        Page page = userService.queryPage(params);
        return Result.ok().put("page", page);
    }

接下对service层的实现,功能是实现关键字搜索，这里因为简单并没有单独将Specification提出来，主要是对Specification接口有个大概的认识。

@Override
    public page queryPage(Map<String, Object> params) {
        //MapUtils方法用来取除params中的方法，来自于 org.apache.commons.collections.MapUtils;
        String keyword = MapUtils.getSrting(params,"keyword");
        Page page = sysUserRepository.findAll(new Specification<SysUserEntity>() {
            @Override
            public Predicate toPredicate(Root<SysUserEntity> root, CriteriaQuery<?> criteriaQuery, CriteriaBuilder criteriaBuilder) {
                List<Predicate> predicates = new ArrayList<>();
                if (StringUtils.isNotBlank(keyword)) {
                    List<Predicate> temp = new ArrayList<>();
                    for (String oneKeyword : keywordCopyStr) {
                        temp.add(criteriaBuilder.like(root.<String>get("mobile"), "%" + oneKeyword + "%"));
                        temp.add(criteriaBuilder.like(root.<String>get("trueName"), "%" + oneKeyword + "%"));
                    }
                    predicates.add(criteriaBuilder.or(temp.toArray(new Predicate[temp.size()])));
                }
                return criteriaQuery.where(predicates.toArray(new Predicate[predicates.size()])).getRestriction();
            }
        });
        return page;
    }

Repository层中为了支持这样的查询，sysUserRepository需要继承JpaRepository（基本查询），JpaSpecificationExecutor（分页），这个接口是不需要再去实现的，到了Repository层就行，再对此进行扩充（比Mybatis简单多了）。

public interface sysUserRepository extends JpaRepository<SysCaptchaEntity, String>, JpaSpecificationExecutor<SysCaptchaEntity> {

}

Repository层常用写法

按着类字段查询按着类关系关联查询不需要写语句的查询就不在详细讲解，给个例子看看就行

//查询第一个
EngineerVersionControl findTopByProjectIdAndOldOrderByVersionDesc(String projectId, int old);
//查询存在
boolean existsByProjectId(String projectId);
//排序
List<EngineerVersionControl> findByProjectIdOrderByVersionDesc(String projectId);

@Query语句查询

一般不会做物理删除，而是逻辑删除。保存操作使用save或者saveAll方法

//更新 必有@Modifying，和使用hiberna一样，HQL语句的写法
@Transactional(rollbackFor = Exception.class)
@Modifying
@Query("update SysUserEntity  u set u.isDelete = ?2 ,u.gmtModified= ?3  where u.id = ?1 ")
int deleteIsUpdate(String id, int deleted, Date date);

//查询 
@Query("select u from SysUserEntity u where  u.id in ?1 and u.isDelete = 0 ")
Page<SysUserEntity> findAllUser(List<String> userIds, Pageable pageable);

//多表查询
 @Query(value = " select p  from DictionaryEntity p , DictionaryContentEntity w " +
            " Where w.ContentEntity.id = ?1 and p.id = w.DictionaryEntity.id and p.deleted = ?2 ORDER BY p.dictionary")
    List<WebsiteDictionaryEntity> webOnwDictionary(String id,int isDeleted);

3.使用@Query实现写sql语句的查询
再spring data 中不仅有HQl语句，在功能太复杂的时候，可以使用sql语句进行本地查询

  @Query(value="select serve.* from service_serve serve " +
            "left join company_info_user cominfo on serve.company_info_user_id=cominfo.company_info_user_id" +
            " left  join user_company company on cominfo.company_info_user_id=company.company_info_user_id " +
            " left  join employee employeeen3_ on company.user_company_id=employeeen3_.user_company_id " +
            " left  join user userentity4_  on employeeen3_.user_id=userentity4_.user_id " +
            " where userentity4_.user_id=?1 and employeeen3_.activity_management_power=1" +
            " order by serve.gmt_create desc " +
            "limit ?2 , ?3 " ,nativeQuery = true)
    List<ServiceServeEntity> queryByCompany(String userId,int startPoint,int endPoint);

4.@Param（value = “name”）查询，这是两种写法。一种?，一种@Param()

 @Query(value="select activity.* from service_activity activity " +
            "left join company_info_user cominfo on activity.company_id=cominfo.company_info_user_id" +
            " left  join user_company company on cominfo.company_info_user_id=company.company_info_user_id " +
            " left  join employee employeeen3_ on company.user_company_id=employeeen3_.user_company_id " +
            " left  join user userentity4_  on employeeen3_.user_id=userentity4_.user_id " +
            " where userentity4_.user_id=:userId and employeeen3_.service_management_power=1 and activity.status=:status" +
            " order by activity.gmt_create desc " +
            "limit :startPoint , :endPoint " ,nativeQuery = true)
    List<ServiceActivityEntity> queryByCompanyAndStatus(@Param(value="userId")String userId,@Param(value="status")int status, @Param(value="startPoint")int startPoint, @Param(value="endPoint")int endPoint);

Specification 的用法

下面是个较为全面的例子，将一个较为复杂的查询提取成一个方法。这个方法时使用and的方式拼接，接下来的每一个查询都需要使用把finalConditions拼上，如同 finalConditions = criteriaBuilder.and(finalConditions, taskFastPre)。

public class TaskProjectSpecs {
    public static Specification<Task> where(Map params, String userId, List<String> taskIds) {
        //lambda表达式
        return (Root<Task> root, CriteriaQuery<?> query, CriteriaBuilder criteriaBuilder) -> {
            //开始
            Predicate finalConditions = criteriaBuilder.conjunction();
            
            //提取参数
            String taskFast = MapUtils.getString(params, "taskFast");

            //lile 和join 用法 join可跟,JoinType.LEFT等
            if (StringUtils.isNotBlank(taskFast)) {
                Predicate taskFastPre = criteriaBuilder.like(root.join("taskType",JoinType.LEFT).<String>get("id"), "%" + taskFast + "%");
                finalConditions = criteriaBuilder.and(finalConditions, taskFastPre);
            }
            //between用法
            if ((null != createBegin) && (null != createEnd)) {
                Predicate datePredicate = null;
                if (createBegin.after(createEnd)) {
                    datePredicate = criteriaBuilder.between(root.get("gmtCreate"), createEnd, createBegin);
                } else {
                    datePredicate = criteriaBuilder.between(root.get("gmtCreate"), createBegin, createEnd);
                }
                finalConditions = criteriaBuilder.and(finalConditions, datePredicate);
            }
            //equale
            if (null != emergency && 0 != emergency) {
                finalConditions = criteriaBuilder.and(finalConditions, criteriaBuilder.equal(root.get("emergencyLevel"), emergency));
            }
            //大于 不等于
            if (status != null) {
                finalConditions = criteriaBuilder.and(finalConditions, criteriaBuilder.greaterThan(root.get("startDate"), new Date()));
                finalConditions = criteriaBuilder.and(finalConditions, criteriaBuilder.notEqual(root.get("status"), 1));
                
            }
            // or 
            if (StringUtils.isNotBlank(keyword)) {
                finalConditions = criteriaBuilder.and(finalConditions, criteriaBuilder.or(
                        criteriaBuilder.like(root.get("taskName"), "%" + keyword + "%"),
                        criteriaBuilder.like(root.join("project").get("name"), "%" + keyword + "%"))
                );
            }
            //in
            if (taskIds.size() > 0) {
                CriteriaBuilder.In<Object> in = criteriaBuilder.in(root.get("id"));
                for (String id : taskIds) {
                    in.value(id);
                }
                finalConditions = criteriaBuilder.and(finalConditions, in);
            }
            return query.where(finalConditions).getRestriction();
        };
    }
}

上面的方法是and凭借，还有一种add的方法，本质一样，都是构建query.where（）查询。

public class UserSpecs {
    public static Specification<SysUserEntity> where(String keyword, Date createdAtBegin, Date createdAtEnd, List<String> userIds) {
        return (Root<SysUserEntity> root, CriteriaQuery<?> query, CriteriaBuilder cb) -> {
            List<Predicate> predicates = new ArrayList<>();
            if (StringUtils.isNotBlank(keyword)) {
                List<Predicate> temp = new ArrayList<>();
                Set<String> keywordCopyStr = StringUtil.cutToArray(keyword);
                for (String oneKeyword : keywordCopyStr) {
                    temp.add(cb.like(root.<String>get("mobile"), "%" + oneKeyword + "%"));
                    temp.add(cb.like(root.<String>get("trueName"), "%" + oneKeyword + "%"));
                }
                predicates.add(cb.or(temp.toArray(new Predicate[temp.size()])));
            }
            //未删除
            predicates.add(cb.equal(root.get("isDelete"), Constant.NOT_DELETED));
            query.where(predicates.toArray(new Predicate[predicates.size()]));
            return query.getRestriction();
        };
    }
}
return query.where(predicates.toArray(new Predicate[predicates.size()])).getRestriction();

总结

基本将常用的都包含完了，以后遇到了新的写法再更新上去。