使用spring-data-jpa和spring-mvc过滤数据库行

Question

我有一个spring-mvc项目,它使用spring-data-jpa进行数据访问.我有一个域对象Travel,我想让最终用户对它应用一些过滤器.

为此,我实现了以下控制器:

@Autowired
private TravelRepository travelRep;

@RequestMapping("/search")  
public ModelAndView search(
        @RequestParam(required= false, defaultValue="") String lastName, 
        Pageable pageable) {  
    ModelAndView mav = new ModelAndView("travels/list");  
    Page<Travel> travels  = travelRep.findByLastNameLike("%"+lastName+"%", pageable);
    PageWrapper<Travel> page = new PageWrapper<Travel>(travels, "/search");
    mav.addObject("page", page);
    mav.addObject("lastName", lastName);
    return mav;
}

这很好用:用户有一个带有lastName输入框的表单,可以用来过滤旅行.

除了lastName之外,我的Travel域对象还有很多我想要过滤的属性.我认为如果这些属性都是字符串,那么我可以将它们添加为@RequestParams并添加一个spring-data-jpa方法来进行查询.例如,我添加了一个方法findByLastNameLikeAndFirstNameLikeAndShipNameLike.

但是,当我需要过滤外键时,我不知道该怎么做.所以我Travel有一个period属性是Period域对象的外键,我需要将它作为下拉列表供用户选择Period.

我想要做的是当句点为空时我想要检索由lastName过滤的所有旅行,当句点不为空时我想要检索由lastName过滤的这段时间的所有旅行.

我知道如果我在我的存储库中实现两个方法并使用一个if到我的控制器,这可以做到:

public ModelAndView search(
       @RequestParam(required= false, defaultValue="") String lastName,
       @RequestParam(required= false, defaultValue=null) Period period, 
       Pageable pageable) {  
  ModelAndView mav = new ModelAndView("travels/list");  
  Page travels = null;
  if(period==null) {
    travels  = travelRep.findByLastNameLike("%"+lastName+"%", pageable);
  } else {
    travels  = travelRep.findByPeriodAndLastNameLike(period,"%"+lastName+"%", pageable);
  }
  mav.addObject("page", page);
  mav.addObject("period", period);
  mav.addObject("lastName", lastName);
  return mav;
}

有没有办法这样做而不使用if？我的旅行不仅有时间段,还有其他需要使用下拉菜单过滤的属性!您可以理解,当我需要使用更多下拉菜单时,复杂性会呈指数级增加,因为需要考虑所有组合:(

更新03/12/13:继续M. Deinum的优秀答案,并在实际实施之后,我想就问题的完整性提供一些意见/ asnwer:

1. 而不是实现JpaSpecificationExecutor您应该实现JpaSpecificationExecutor<Travel>以避免类型检查警告.

2. 请看看kostja出色的回答这个问题 真的动态JPA CriteriaBuilder ,因为你将需要实现这一点,如果你想拥有正确的过滤器.

3. 我能够为Criteria API找到的最佳文档是http://www.ibm.com/developerworks/library/j-typesafejpa/.这是一个相当长的阅读,但我完全推荐它 - 在阅读之后,我对Root和CriteriaBuilder的大部分问题都得到了回答:)

4. 重用该Travel对象是不可能的,因为它包含我需要搜索的各种其他对象(也包含其他对象)Like- 而是使用了TravelSearch包含我需要搜索的字段的对象.

2015年5月10日更新:根据@ priyank的请求,这是我实现TravelSearch对象的方式:

public class TravelSearch {
    private String lastName;
    private School school;
    private Period period;
    private String companyName;
    private TravelTypeEnum travelType;
    private TravelStatusEnum travelStatus;
    // Setters + Getters
}

TravelSpecification使用了这个对象(大部分代码都是特定于域的,但我将其留在那里作为例子):

public class TravelSpecification implements Specification<Travel> {
    private TravelSearch criteria;


    public TravelSpecification(TravelSearch ts) {
        criteria= ts;
    }

    @Override
    public Predicate toPredicate(Root<Travel> root, CriteriaQuery<?> query, 
            CriteriaBuilder cb) {
        Join<Travel, Candidacy> o = root.join(Travel_.candidacy);

        Path<Candidacy> candidacy = root.get(Travel_.candidacy);
        Path<Student> student = candidacy.get(Candidacy_.student);
        Path<String> lastName = student.get(Student_.lastName);
        Path<School> school = student.get(Student_.school);

        Path<Period> period = candidacy.get(Candidacy_.period);
        Path<TravelStatusEnum> travelStatus = root.get(Travel_.travelStatus);
        Path<TravelTypeEnum> travelType = root.get(Travel_.travelType);

        Path<Company> company = root.get(Travel_.company);
        Path<String> companyName = company.get(Company_.name);

        final List<Predicate> predicates = new ArrayList<Predicate>();
        if(criteria.getSchool()!=null) {
            predicates.add(cb.equal(school, criteria.getSchool()));
        }
        if(criteria.getCompanyName()!=null) {
            predicates.add(cb.like(companyName, "%"+criteria.getCompanyName()+"%"));
        }
        if(criteria.getPeriod()!=null) {
            predicates.add(cb.equal(period, criteria.getPeriod()));
        }
        if(criteria.getTravelStatus()!=null) {
            predicates.add(cb.equal(travelStatus, criteria.getTravelStatus()));
        }
        if(criteria.getTravelType()!=null) {
            predicates.add(cb.equal(travelType, criteria.getTravelType()));
        }
        if(criteria.getLastName()!=null ) {
            predicates.add(cb.like(lastName, "%"+criteria.getLastName()+"%"));
        }
        return cb.and(predicates.toArray(new Predicate[predicates.size()]));

    }
}

最后,这是我的搜索方法:

@RequestMapping("/search")  
public ModelAndView search(
        @ModelAttribute TravelSearch travelSearch,
        Pageable pageable) {  
    ModelAndView mav = new ModelAndView("travels/list");  

    TravelSpecification tspec = new TravelSpecification(travelSearch);

    Page<Travel> travels  = travelRep.findAll(tspec, pageable);

    PageWrapper<Travel> page = new PageWrapper<Travel>(travels, "/search");

    mav.addObject(travelSearch);

    mav.addObject("page", page);
    mav.addObject("schools", schoolRep.findAll() );
    mav.addObject("periods", periodRep.findAll() );
    mav.addObject("travelTypes", TravelTypeEnum.values());
    mav.addObject("travelStatuses", TravelStatusEnum.values());
    return mav;
}

希望我帮忙!

Answer 1

对于初学者,你应该停止使用@RequestParam并将所有搜索字段放在一个对象中(可能会重复使用Travel对象).然后,您有2个选项可用于动态构建查询

1. 使用JpaSpecificationExecutor和写一个Specification
2. 使用QueryDslPredicateExecutor并使用QueryDSL编写谓词.

运用 JpaSpecificationExecutor

首先添加JpaSpecificationExecutor到您的TravelRepository这将为您提供一种findAll(Specification)方法,您可以删除自定义查找器方法.

public interface TravelRepository extends JpaRepository<Travel, Long>, JpaSpecificationExecutor<Travel> {}

然后,您可以在存储库中创建一个方法,该方法使用Specification基本构建查询的方法.有关此信息,请参阅Spring Data JPA 文档.

您唯一需要做的就是创建一个实现的类,Specification并根据可用的字段构建查询.使用JPA Criteria API链接构建查询.

public class TravelSpecification implements Specification<Travel> {

    private final Travel criteria;

    public TravelSpecification(Travel criteria) {
        this.criteria=criteria;
    }

    public Predicate toPredicate(Root<T> root, CriteriaQuery<?> query, CriteriaBuilder builder) {
        // create query/predicate here.
    }
}

最后你需要修改你的控制器以使用新findAll方法(我冒昧地清理它一点).

@RequestMapping("/search")  
public String search(@ModelAttribute Travel search, Pageable pageable, Model model) {  
Specification<Travel> spec = new TravelSpecification(search);
    Page<Travel> travels  = travelRep.findAll(spec, pageable);
    model.addObject("page", new PageWrapper(travels, "/search"));
    return "travels/list";
}

运用 QueryDslPredicateExecutor

首先添加QueryDslPredicateExecutor到您的TravelRepository这将为您提供一种findAll(Predicate)方法,您可以删除自定义查找器方法.

public interface TravelRepository extends JpaRepository<Travel, Long>, QueryDslPredicateExecutor<Travel> {}

接下来,您将实现一个服务方法,该方法将使用该Travel对象使用QueryDSL构建谓词.

@Service
@Transactional
public class TravelService {

    private final TravelRepository travels;

    public TravelService(TravelRepository travels) {
        this.travels=travels;
    }

    public Iterable<Travel> search(Travel criteria) {

        BooleanExpression predicate = QTravel.travel...
        return travels.findAll(predicate);
    }
}

另见这个沼泽的帖子.