StringBuffer queryString=new StringBuffer();
  queryString.append("select distinct purchase from ConPurchaseInfo purchase,ConSalesInfo sales where sales.salesContractno    like '%"+value+"%' and purchase in elements(sales.conPurchaseInfo) and (purchase.id in(")
             .append(contractIds)
             .append(") or purchase.createUserId='"+userInfo.get(Constant.USERID_SESSION)+"' ) and purchase.recycleFlag=0 and purchase.contractStatus=0");
     if("user".equals(isUserOrOrg)){
      queryString.append(" and purchase.purchaseChiefId='"+id+"' ");
     }else if("org".equals(isUserOrOrg)){
      queryString.append(" and purchase.purchaseChiefDepartmentId in("+getOrgStrs(id,true)+") ");
     }else{
      queryString.append(" ");
     }
   if(StringUtils.isNotBlank(state)){
    queryString.append(" and ");
    queryString.append("purchase.applicationStatus ='" + state + "'");
   }
   queryString.append(" order by purchase.updateTime desc,purchase.signedDate desc");
 

是否有過這種需求？物件之間是多對多，需要一方查詢另一方，而條件是一方除了滿足自身的where限制意外，還要受關聯的多的一方的限制。

比如：

public class ConPurchaseInfo extends BaseEntityUUID implements ResourceEntity{
	private Set<ConSalesInfo> conSalesInfos;
	private static final long serialVersionUID = 1L;｝

ConPurchaseInfo 中有對ConSalesInfo的關聯many-to-many，

現在要根據conSalesInfo中的屬性name來查詢conPurchaseInfo的資訊。

用sql是可以寫的，現在介紹hql寫法。

用到了in 和 in elements這兩個不同的用法

其中in 都知道了把，in elements 是專門針對set集合的。

最簡單例子hql如下：

select distinct purchase from ConPurchaseInfo purchase,ConSalesInfo sales where sales.name like '%"+value+"%' and purchasein elements(sales.conPurchaseInfo)；