grails - 使用joinTable进行一对多映射

Question

我有两个域类。一个是“合作伙伴”，另一个是“客户”。客户可以是合作伙伴的一部分，并且合作伙伴可以拥有1个或多个客户:

class Customer {
    Integer id
    String name
    static hasOne = [partner:Partner]
    static mapping = {
        partner joinTable:[name:'PartnerMap',column:'partner_id',key:'customer_id']
    }
}

class Partner {
    Integer id
    static hasMany = [customers:Customer]
    static mapping = {
        customers joinTable:[name:'PartnerMap',column:'customer_id',key:'partner_id']
    }
}

但是，每当我尝试查看客户是否是合作伙伴的一部分时，都像这样:

def customers = Customer.list()
customers.each {
     if (it.partner) {
          println "Partner!"
     }
}

我收到以下错误:

org.springframework.dao.InvalidDataAccessResourceUsageException: could not execute query; SQL [select this_.customer_id as customer1_162_0_, this_.company as company162_0_, this_.display_name as display3_162_0_, this_.parent_customer_id as parent4_162_0_, this_.partner_id as partner5_162_0_, this_.server_id as server6_162_0_, this_.status as status162_0_, this_.vertical_market as vertical8_162_0_ from Customer this_]; nested exception is org.hibernate.exception.SQLGrammarException: could not execute query

看起来Grails好像在认为partner_id是Customer查询的一部分，而不是……它在PartnerMap表中，该表应该找到customer_id，然后从相应的partner_id中获取Partner。

有人知道我在做什么错吗？

编辑:我忘了提到我正在使用旧数据库表执行此操作。因此，我有一个Partner，Customer和PartnerMap表。 PartnerMap仅具有一个customer_id和partner_id字段。

Answer 1

考虑到当您想要一个联接表时“一对多”工作的方式，我认为使用标准GORM不可能使其成为双向并访问客户的合作伙伴。但是您可以将联接表与域类映射，并以这种方式访问​​内容:

顾客:

class Customer {
   String name
   def getPartner() {
      PartnerMap.findByCustomer(this)?.partner
   }
}

伙伴:

class Partner {
   String name
   def getCustomers() {
      PartnerMap.findAllByPartner(this)*.customer
   }
}

合作伙伴 map :

import org.apache.commons.lang.builder.HashCodeBuilder

class PartnerMap implements Serializable {

   Partner partner
   Customer customer

   boolean equals(other) {
      if (!(other instanceof PartnerMap)) {
         return false
      }

      other.partner?.id == partner?.id &&
         other.customer?.id == customer?.id
   }

   int hashCode() {
      def builder = new HashCodeBuilder()
      if (partner) builder.append(partner.id)
      if (customer) builder.append(customer.id)
      builder.toHashCode()
   }

   static PartnerMap get(long partnerId, long customerId) {
      find 'from PartnerMap where partner.id=:partnerId and customer.id=:customerId',
         [partnerId: partnerId, customerId: customerId]
   }

   static PartnerMap create(Partner partner, Customer customer, boolean flush = false) {
      new PartnerMap(partner: partner, customer: customer).save(flush: flush, insert: true)
   }

   static boolean remove(Partner partner, Customer customer, boolean flush = false) {
      PartnerMap instance = PartnerMap.findByPartnerAndCustomer(partner, customer)
      instance ? instance.delete(flush: flush) : false
   }

   static void removeAll(Partner partner) {
      executeUpdate 'DELETE FROM PartnerMap WHERE partner=:partner', [partner: partner]
   }

   static void removeAll(Customer customer) {
      executeUpdate 'DELETE FROM PartnerMap WHERE customer=:customer', [customer: customer]
   }

   static mapping = {
      id composite: ['customer', 'partner']
      version false
      table 'PartnerMap'
   }
}

由于未使用hasMany，因此会丢失addToXXX动态方法，但是可以调用 PartnerMap.create()来关联两个实例。您还会丢失域类中的集合和反向引用，但是我为它们添加了实用程序方法。