hibernate根据条件动态组装sql/hql语句(仿ibatis动态查询语句功

1.功能需求背景
项目中使用hibernate作为数据持久层框架，主要考虑hibernate在进行一些简单的crud操作时非常便利，不需要和ibatis似的为每个sql操作都写一堆文件，但是同时也带来了一些局限性，如类似ibatis强大的动态查询功能用不了了，但是这个功能在项目中的应用场景又很大，hibernate自身也支持将sql/hql语句写在.hbm.xml映射文件中<sql-query>和<query>元素，但是这个功能只能对那些查询限制条件固定的sql有用，对于需要动态拼接的sql语句，hibernate就显得力不从心了，如何给hibernate插上ibatis动态查询的翅膀，既保留crud的简洁性，又能收获ibatis的特性呢？接下来的文章将会重点介绍
2.设计思路
先看一下ibatis的动态查询时怎么做的
<select id="getUserList" resultMap="user">
select * from user
<isGreaterThan prepend="and" property="id" compareValue="0">
where user_id = #userId#
</isGreaterThan>
order by createTime desc
</select>
ibatis在程序实现内部回去解析sql语句中的标签，然后去解析计算，我们在ibatis在实现的时候也参考了这个解决思路，但是否是需要把ibatis里的解析sql的语法都抄到我们的dao框架中呢-显然这样太复杂了，而且ibatis自己的sql元素是和那些resultMap等是绑定在一起用的，而在hibernate是没用这些东西的，要改造这些东西是一项非常浩大的工程，因此这个方案被放弃了

我们在实现的时候采取了一种非常简洁又功能强大的方式-模板技术!对，就是利用freemarker把sql/hql中的动态拼接条件判断语法都交给freemarker语法去处理，这样既能复用freemarker框架，又保持了我们框架设计的简洁性-不需要自己写过多的处理逻辑,以下是我们需要进行动态处理的sql/hql语句的样例

<?xml version="1.0" encoding="utf-8"?><!DOCTYPE dynamic-hibernate-statement PUBLIC "-//Haier/HOP Hibernate Dynamic Statement DTD 1.0//EN""http://www.haier.com/dtd/dynamic-hibernate-statement-1.0.dtd"><dynamic-hibernate-statement><!-- 查询某个资源下的直接子节点 --><hql-query name="resource.getChildren"><![CDATA[from Resource where parent.id=${parentId} and parent.id != id]]> </hql-query><!-- 查询系统中所有的root资源 --><hql-query name="resource.getRoots"><![CDATA[from Resource where parent.id = id order by orderIndex]]> </hql-query><!-- 获取某个用户可访问的某个资源下的所有子资源 --><sql-query name="resource.getDescendants"><![CDATA[select distinct t.id,                t.name,                t.description,                t.url,                t.type,                t.status,                t.code,                t.configuration,                t.module_name,                t.gmt_create,                t.gmt_modified,                t.create_by,                t.last_modified_by,                t.order_index,                t.parent_id  from resource_info t inner join role_resource rr    on t.id = rr.resource_id inner join user_role ur    on rr.role_id = ur.role_id where ur.user_id = ${userId} <#if type == '1'> and t.type=1 <#else> and t.type=0 </#if>   and t.type =  ${type}   and t.status =  ${status} start with t.code = '${code}'connect by nocycle prior t.id = t.parent_id]]> </sql-query></dynamic-hibernate-statement>

这个文件看起来非常类似ibatis的语句了，只是没用ibatis的哪些标签-改成了freemarker语法，没错，我们就是复用freemarker来帮我们解决这些烦杂的判断操作的

这样我们的动态sql程序就可以总结成以下流程

a.系统加载阶段

这个阶段程序负责将指定路径下的动态sql文件加载到内存中，一次性缓存起来，没错，这些东西只需要加载一次，以后直接读取就行了，没必要每次去查找，缓存也非常简单，一个Map<String,String>就搞定，key是sql-query或hql-query元素的name属性，value就是与其对应的sql/hql语句

b.程序调用查询阶段

调用程序通过sql/hql语句的name属性和传入查询参数来得到最终解析出来的语句

我们期望的方法可能是这样的：

public <X> List<X> findByNamedQuery(final String queryName, final Map<String, ?> parameters) 

通过queryName从缓存中查找出其对应的sql/hql语句(最原始的，里面带有freemarker语法)

然后通过freemarker模板和传递进去的parameters参数对模板进行解析，得到最终的语句(纯sql/hql)

最后将解析后的sql/hql传递给底层api，返回查询结果

3.实现

上面介绍了大致的思路，这里介绍具体的代码实现

3.1DTD定义

我们是把动态的sql/hql语句放在单独的xml配置文件里的，为了规范xml文档，我们给文档定义了dtd文件，这里我们只定义了两个元素<sql-query>和<hql-query>分别表示sql查询语句和hql查询语句，这两个元素目前自有一个name属性用来唯一标示该语句，如下

<!-- HOP Hibernate Dynamic Statement Mapping DTD.<!DOCTYPE dynamic-hibernate-statement PUBLIC     "-//Haier/HOP Hibernate Dynamic Statement DTD 1.0//EN"    "http://www.haier.com/dtd/dynamic-hibernate-statement-1.0.dtd">这个文件时用来定义动态参数语句，类似itabis--><!--The document root. --><!ELEMENT dynamic-hibernate-statement ((hql-query|sql-query)*)><!-- default: none --><!-- The query element declares a named Hibernate query string --><!ELEMENT hql-query (#PCDATA)><!ATTLIST hql-query name CDATA #REQUIRED><!-- The sql-query element declares a named SQL query string --><!ELEMENT sql-query (#PCDATA)><!ATTLIST sql-query name CDATA #REQUIRED>

然后将其保存为dynamic-hibernate-statement-1.0.dtd，放在classpath下

编写DTD校验器

/** * hibernate动态sql dtd解析器 * @author WangXuzheng * */public class DynamicStatementDTDEntityResolver implements EntityResolver, Serializable{private static final long serialVersionUID = 8123799007554762965L;private static final Logger LOGGER = LoggerFactory.getLogger( DynamicStatementDTDEntityResolver.class );private static final String HOP_DYNAMIC_STATEMENT = "http://www.haier.com/dtd/";public InputSource resolveEntity(String publicId, String systemId) {InputSource source = null; // returning null triggers default behaviorif ( systemId != null ) {LOGGER.debug( "trying to resolve system-id [" + systemId + "]" );if ( systemId.startsWith( HOP_DYNAMIC_STATEMENT ) ) {LOGGER.debug( "recognized hop dyanmic statement namespace; attempting to resolve on classpath under com/haier/openplatform/dao/hibernate/" );source = resolveOnClassPath( publicId, systemId, HOP_DYNAMIC_STATEMENT );}}return source;}private InputSource resolveOnClassPath(String publicId, String systemId, String namespace) {InputSource source = null;String path = "com/haier/openplatform/dao/hibernate/" + systemId.substring( namespace.length() );InputStream dtdStream = resolveInHibernateNamespace( path );if ( dtdStream == null ) {LOGGER.debug( "unable to locate [" + systemId + "] on classpath" );if ( systemId.substring( namespace.length() ).indexOf( "2.0" ) > -1 ) {LOGGER.error( "Don't use old DTDs, read the Hibernate 3.x Migration Guide!" );}}else {LOGGER.debug( "located [" + systemId + "] in classpath" );source = new InputSource( dtdStream );source.setPublicId( publicId );source.setSystemId( systemId );}return source;}protected InputStream resolveInHibernateNamespace(String path) {return this.getClass().getClassLoader().getResourceAsStream( path );}protected InputStream resolveInLocalNamespace(String path) {try {return ConfigHelper.getUserResourceAsStream( path );}catch ( Throwable t ) {return null;}}}

3.2编写sql文件

<?xml version="1.0" encoding="utf-8"?><!DOCTYPE dynamic-hibernate-statement PUBLIC "-//Haier/HOP Hibernate Dynamic Statement DTD 1.0//EN""http://www.haier.com/dtd/dynamic-hibernate-statement-1.0.dtd"><dynamic-hibernate-statement><!-- 查询某个资源下的直接子节点 --><hql-query name="resource.getChildren"><![CDATA[from Resource where parent.id=${parentId} and parent.id != id]]> </hql-query><!-- 查询系统中所有的root资源 --><hql-query name="resource.getRoots"><![CDATA[from Resource where parent.id = id order by orderIndex]]> </hql-query><!-- 获取某个用户可访问的某个资源下的所有子资源 --><sql-query name="resource.getDescendants"><![CDATA[select distinct t.id,                t.name,                t.description,                t.url,                t.type,                t.status,                t.code,                t.configuration,                t.module_name,                t.gmt_create,                t.gmt_modified,                t.create_by,                t.last_modified_by,                t.order_index,                t.parent_id  from resource_info t inner join role_resource rr    on t.id = rr.resource_id inner join user_role ur    on rr.role_id = ur.role_id where ur.user_id = ${userId} <#if type == '1'> and t.type=1 <#else> and t.type=0 </#if>   and t.type =  ${type}   and t.status =  ${status} start with t.code = '${code}'connect by nocycle prior t.id = t.parent_id]]> </sql-query></dynamic-hibernate-statement>

3.3加载动态sql文件

这里我们将加载sql/hql语句的程序独立到一个单独的类中，以便独立扩展

这里一共3个方法，分表标示获取系统中sql/hql语句的map(key:语句名称,value:具体的)

/** * 动态sql/hql语句组装器 * @author WangXuzheng * */public interface DynamicHibernateStatementBuilder {/** * hql语句map * @return */public Map<String,String> getNamedHQLQueries();/** * sql语句map * @return */public Map<String,String> getNamedSQLQueries();/** * 初始化 * @throws IOException  */public void init() throws IOException;}

默认的加载器-将指定配置文件中的sql/hql语句加载到内存中

/** * @author WangXuzheng * */public class DefaultDynamicHibernateStatementBuilder implements DynamicHibernateStatementBuilder, ResourceLoaderAware {private static final Logger LOGGER = LoggerFactory.getLogger(DefaultDynamicHibernateStatementBuilder.class);private Map<String, String> namedHQLQueries;private Map<String, String> namedSQLQueries;private String[] fileNames = new String[0];private ResourceLoader resourceLoader;private EntityResolver entityResolver = new DynamicStatementDTDEntityResolver();/** * 查询语句名称缓存，不允许重复 */private Set<String> nameCache = new HashSet<String>();public void setFileNames(String[] fileNames) {this.fileNames = fileNames;}@Overridepublic Map<String, String> getNamedHQLQueries() {return namedHQLQueries;}@Overridepublic Map<String, String> getNamedSQLQueries() {return namedSQLQueries;}@Overridepublic void init() throws IOException {namedHQLQueries = new HashMap<String, String>();namedSQLQueries = new HashMap<String, String>();boolean flag = this.resourceLoader instanceof ResourcePatternResolver;for (String file : fileNames) {if (flag) {Resource[] resources = ((ResourcePatternResolver) this.resourceLoader).getResources(file);buildMap(resources);} else {Resource resource = resourceLoader.getResource(file);buildMap(resource);}}//clear name cachenameCache.clear();}@Overridepublic void setResourceLoader(ResourceLoader resourceLoader) {this.resourceLoader = resourceLoader;}private void buildMap(Resource[] resources) throws IOException {if (resources == null) {return;}for (Resource resource : resources) {buildMap(resource);}}@SuppressWarnings({ "rawtypes" })private void buildMap(Resource resource) {InputSource inputSource = null;try {inputSource = new InputSource(resource.getInputStream());XmlDocument metadataXml = MappingReader.INSTANCE.readMappingDocument(entityResolver, inputSource,new OriginImpl("file", resource.getFilename()));if (isDynamicStatementXml(metadataXml)) {final Document doc = metadataXml.getDocumentTree();final Element dynamicHibernateStatement = doc.getRootElement();Iterator rootChildren = dynamicHibernateStatement.elementIterator();while (rootChildren.hasNext()) {final Element element = (Element) rootChildren.next();final String elementName = element.getName();if ("sql-query".equals(elementName)) {putStatementToCacheMap(resource, element, namedSQLQueries);} else if ("hql-query".equals(elementName)) {putStatementToCacheMap(resource, element, namedHQLQueries);}}}} catch (Exception e) {LOGGER.error(e.toString());throw new SysException(e);} finally {if (inputSource != null && inputSource.getByteStream() != null) {try {inputSource.getByteStream().close();} catch (IOException e) {LOGGER.error(e.toString());throw new SysException(e);}}}}private void putStatementToCacheMap(Resource resource, final Element element, Map<String, String> statementMap)throws IOException {String sqlQueryName = element.attribute("name").getText();Validate.notEmpty(sqlQueryName);if (nameCache.contains(sqlQueryName)) {throw new SysException("重复的sql-query/hql-query语句定义在文件:" + resource.getURI() + "中，必须保证name的唯一.");}nameCache.add(sqlQueryName);String queryText = element.getText();statementMap.put(sqlQueryName, queryText);}private static boolean isDynamicStatementXml(XmlDocument xmlDocument) {return "dynamic-hibernate-statement".equals(xmlDocument.getDocumentTree().getRootElement().getName());}}

配置一下

<bean id="dynamicStatementBuilder" class="com.haier.openplatform.dao.hibernate.support.DefaultDynamicHibernateStatementBuilder"><property name="fileNames"><list><value>classpath*:/**/*-dynamic.xml</value><!--这里我们指定要加载某个文件夹下所有以-dynamic.xml结尾的文件 --></list></property></bean>    <bean id="baseHibernateDAO" class="com.haier.openplatform.dao.hibernate.BaseDAOHibernateImpl" abstract="true">        <property name="sessionFactory">            <ref bean="sessionFactory"/>        </property>        <property name="dynamicStatementBuilder" ref="dynamicStatementBuilder"/>    </bean>

dao层代码

/** * Hibernate实现的DAO层 * @param <T> DAO操作的对象类型 * @param <ID> 主键类型 * @author WangXuzheng * */public class SimpleHibernateDAO<T,ID extends Serializable> implements BaseDAO<T, ID>,InitializingBean{private static final Logger LOGER = LoggerFactory.getLogger(SimpleHibernateDAO.class);protected SessionFactory sessionFactory;protected Class<T> entityClass;/** * 模板缓存 */protected Map<String, StatementTemplate> templateCache;protected DynamicHibernateStatementBuilder dynamicStatementBuilder;/** * 通过子类的泛型定义取得对象类型Class. * eg. * public class UserDao extends SimpleHibernateDao<User, Long> */public SimpleHibernateDAO() {this.entityClass = Reflections.getSuperClassGenricType(getClass());}/** * 取得sessionFactory. */public SessionFactory getSessionFactory() {return sessionFactory;}/** * 采用@Autowired按类型注入SessionFactory, 当有多个SesionFactory的时候在子类重载本函数. */public void setSessionFactory(final SessionFactory sessionFactory) {this.sessionFactory = sessionFactory;}public void setDynamicStatementBuilder(DynamicHibernateStatementBuilder dynamicStatementBuilder) {this.dynamicStatementBuilder = dynamicStatementBuilder;}/** * 取得当前Session. */public Session getSession() {return sessionFactory.getCurrentSession();}/** * 保存新增或修改的对象. */@Overridepublic void save(final T entity) {Validate.notNull(entity, "entity不能为空");getSession().save(entity);LOGER.debug("save entity: {}", entity);}/** * 删除对象. *  * @param entity 对象必须是session中的对象或含id属性的transient对象. */@Overridepublic void delete(final T entity) {if(entity == null){return;}getSession().delete(entity);LOGER.debug("delete entity: {}", entity);}/** * 按id删除对象. */@Overridepublic void delete(final ID id) {Validate.notNull(id, "id不能为空");delete(get(id));LOGER.debug("delete entity {},id is {}", entityClass.getSimpleName(), id);}/** * 按id获取对象. */@SuppressWarnings("unchecked")@Overridepublic T get(final ID id) {Validate.notNull(id, "id不能为空");return (T) getSession().get(entityClass, id);}/** * 按id列表获取对象列表. */public List<T> get(final Collection<ID> ids) {return find(Restrictions.in(getIdName(), ids));}/** *获取全部对象. */@Overridepublic List<T> getAll() {return find();}/** *获取全部对象, 支持按属性行序. */@SuppressWarnings("unchecked")public List<T> getAll(String orderByProperty, boolean isAsc) {Criteria c = createCriteria();if (isAsc) {c.addOrder(Order.asc(orderByProperty));} else {c.addOrder(Order.desc(orderByProperty));}return c.list();}/** * 按属性查找对象列表, 匹配方式为相等. */public List<T> findBy(final String propertyName, final Object value) {Criterion criterion = Restrictions.eq(propertyName, value);return find(criterion);}/** * 按属性查找唯一对象, 匹配方式为相等. */@SuppressWarnings("unchecked")@Overridepublic T findUniqueBy(final String propertyName, final Object value) {Criterion criterion = Restrictions.eq(propertyName, value);return ((T) createCriteria(criterion).uniqueResult());}/** * 按HQL查询对象列表. *  * @param values 数量可变的参数,按顺序绑定. */@SuppressWarnings("unchecked")public <X> List<X> findByHQL(final String hql, final Object... values) {return createHQLQuery(hql, values).list();}/** * 按HQL查询对象列表，并将对象封装成指定的对象 *  * @param values 数量可变的参数,按顺序绑定. */@SuppressWarnings("unchecked")public <X> List<X> findByHQLRowMapper(RowMapper<X> rowMapper,final String hql, final Object... values) {Validate.notNull(rowMapper, "rowMapper不能为空！");List<Object[]> result = createHQLQuery(hql, values).list();return buildListResultFromRowMapper(rowMapper, result);}protected <X> List<X> buildListResultFromRowMapper(RowMapper<X> rowMapper, List<Object[]> result) {List<X> rs = new ArrayList<X>(result.size());for(Object[] obj : result){rs.add(rowMapper.fromColumn(obj));}return rs;}/** * 按SQL查询对象列表. *  * @param values 数量可变的参数,按顺序绑定. */@SuppressWarnings("unchecked")public <X> List<X> findBySQLRowMapper(RowMapper<X> rowMapper,final String sql, final Object... values) {Validate.notNull(rowMapper, "rowMapper不能为空！");List<Object[]> result = createSQLQuery(sql, values).list();return buildListResultFromRowMapper(rowMapper, result);}/** * 按SQL查询对象列表,并将结果集转换成指定的对象列表 *  * @param values 数量可变的参数,按顺序绑定. */@SuppressWarnings("unchecked")public <X> List<X> findBySQL(final String sql, final Object... values) {return createSQLQuery(sql, values).list();}/** * 按HQL查询对象列表. *  * @param values 命名参数,按名称绑定. */@SuppressWarnings("unchecked")public <X> List<X> findByHQL(final String hql, final Map<String, ?> values) {return createHQLQuery(hql, values).list();}/** * 按HQL查询对象列表,并将结果集封装成对象列表 *  * @param values 命名参数,按名称绑定. */@SuppressWarnings("unchecked")public <X> List<X> findByHQLRowMapper(RowMapper<X> rowMapper,final String hql, final Map<String, ?> values) {Validate.notNull(rowMapper, "rowMapper不能为空！");List<Object[]> result = createHQLQuery(hql, values).list();return buildListResultFromRowMapper(rowMapper, result);}/** * 按SQL查询对象列表. * @param sql SQL查询语句 * @param values 命名参数,按名称绑定. */@SuppressWarnings("unchecked")public <X> List<X> findBySQL(final String sql, final Map<String, ?> values) {return createSQLQuery(sql, values).list();}/** * 查询在xxx.hbm.xml中配置的查询语句 * @param queryName 查询的名称 * @param parameters 参数 * @return */public <X> List<X> findByNamedQuery(final String queryName, final Map<String, ?> parameters) {StatementTemplate statementTemplate = templateCache.get(queryName);String statement = processTemplate(statementTemplate,parameters);if(statementTemplate.getType() == StatementTemplate.TYPE.HQL){return this.findByHQL(statement);}else{return this.findBySQL(statement);}}/** * 查询在xxx.hbm.xml中配置的查询语句 * @param rowMapper * @param queryName 查询的名称 * @param parameters 参数 * @return */public <X> List<X> findByNamedQuery(RowMapper<X> rowMapper,final String queryName, final Map<String, ?> parameters) {StatementTemplate statementTemplate = templateCache.get(queryName);String statement = processTemplate(statementTemplate,parameters);if(statementTemplate.getType() == StatementTemplate.TYPE.HQL){return this.findByHQLRowMapper(rowMapper,statement);}else{return this.findBySQLRowMapper(rowMapper,statement);}}/** * 按SQL查询对象列表,并将结果集封装成对象列表 * @param sql SQL查询语句 * @param values 命名参数,按名称绑定. */@SuppressWarnings("unchecked")public <X> List<X> findBySQLRowMapper(RowMapper<X> rowMapper,final String sql, final Map<String, ?> values) {Validate.notNull(rowMapper, "rowMapper不能为空！");List<Object[]> result = createSQLQuery(sql, values).list();return buildListResultFromRowMapper(rowMapper, result);}/** * 按HQL查询唯一对象. *  * @param values 数量可变的参数,按顺序绑定. */@SuppressWarnings("unchecked")public <X> X findUniqueByHQL(final String hql, final Object... values) {return (X) createHQLQuery(hql, values).uniqueResult();}/** * 按SQL查询唯一对象. *  * @param values 数量可变的参数,按顺序绑定. */@SuppressWarnings("unchecked")public <X> X findUniqueBySQL(final String sql, final Object... values) {return (X) createSQLQuery(sql, values).uniqueResult();}/** * 按HQL查询唯一对象. *  * @param values 命名参数,按名称绑定. */@SuppressWarnings("unchecked")public <X> X findUniqueByHQL(final String hql, final Map<String, ?> values) {return (X) createHQLQuery(hql, values).uniqueResult();}/** * 按HQL查询唯一对象. * @param sql sql语句 * @param values 命名参数,按名称绑定. */@SuppressWarnings("unchecked")public <X> X findUniqueBySQL(final String sql, final Map<String, ?> values) {return (X) createSQLQuery(sql, values).uniqueResult();}/** * 执行HQL进行批量修改/删除操作. *  * @param values 数量可变的参数,按顺序绑定. * @return 更新记录数. */public int batchExecuteHQL(final String hql, final Object... values) {return createHQLQuery(hql, values).executeUpdate();}/** * 执行SQL进行批量修改/删除操作. *  * @param sql sql语句 * @param values 数量可变的参数,按顺序绑定. * @return 更新记录数. */public int batchExecuteSQL(final String sql, final Object... values) {return createSQLQuery(sql, values).executeUpdate();}/** * 执行HQL进行批量修改/删除操作. *  * @param values 命名参数,按名称绑定. * @return 更新记录数. */public int batchExecuteHQL(final String hql, final Map<String, ?> values) {return createHQLQuery(hql, values).executeUpdate();}/** * 执行SQL进行批量修改/删除操作. *  * @param values 命名参数,按名称绑定. * @return 更新记录数. */public int batchExecuteSQL(final String sql, final Map<String, ?> values) {return createSQLQuery(sql, values).executeUpdate();}/** * 根据查询HQL与参数列表创建Query对象. * 与find()函数可进行更加灵活的操作. *  * @param values 数量可变的参数,按顺序绑定. */public Query createHQLQuery(final String queryString, final Object... values) {Query query = getSession().createQuery(queryString);if (values != null) {for (int i = 0; i < values.length; i++) {query.setParameter(i, values[i]);}}return query;}/** * 根据查询SQL与参数列表创建Query对象. * 与find()函数可进行更加灵活的操作. * @param sqlQueryString sql语句 *  * @param values 数量可变的参数,按顺序绑定. */public Query createSQLQuery(final String sqlQueryString, final Object... values) {Query query = getSession().createSQLQuery(sqlQueryString);if (values != null) {for (int i = 0; i < values.length; i++) {query.setParameter(i, values[i]);}}return query;}/** * 根据查询HQL与参数列表创建Query对象. * 与find()函数可进行更加灵活的操作. *  * @param values 命名参数,按名称绑定. */public Query createHQLQuery(final String queryString, final Map<String, ?> values) {Query query = getSession().createQuery(queryString);if (values != null) {query.setProperties(values);}return query;}/** * 根据查询SQL与参数列表创建Query对象. * 与find()函数可进行更加灵活的操作. * @param queryString SQL语句 * @param values 命名参数,按名称绑定. */public Query createSQLQuery(final String queryString, final Map<String, ?> values) {Query query = getSession().createSQLQuery(queryString);if (values != null) {query.setProperties(values);}return query;}/** * 按Criteria查询对象列表. *  * @param criterions 数量可变的Criterion. */@SuppressWarnings("unchecked")public List<T> find(final Criterion... criterions) {return createCriteria(criterions).list();}/** * 按Criteria查询唯一对象. *  * @param criterions 数量可变的Criterion. */@SuppressWarnings("unchecked")public T findUnique(final Criterion... criterions) {return (T) createCriteria(criterions).uniqueResult();}/** * 根据Criterion条件创建Criteria. * 与find()函数可进行更加灵活的操作. *  * @param criterions 数量可变的Criterion. */public Criteria createCriteria(final Criterion... criterions) {Criteria criteria = getSession().createCriteria(entityClass);for (Criterion c : criterions) {criteria.add(c);}return criteria;}/** * 初始化对象. * 使用load()方法得到的仅是对象Proxy, 在传到View层前需要进行初始化. * 如果传入entity, 则只初始化entity的直接属性,但不会初始化延迟加载的关联集合和属性. * 如需初始化关联属性,需执行: * Hibernate.initialize(user.getRoles())，初始化User的直接属性和关联集合. * Hibernate.initialize(user.getDescription())，初始化User的直接属性和延迟加载的Description属性. */public void initProxyObject(Object proxy) {Hibernate.initialize(proxy);}/** * Flush当前Session. */public void flush() {getSession().flush();}/** * 为Query添加distinct transformer. * 预加载关联对象的HQL会引起主对象重复, 需要进行distinct处理. */public Query distinct(Query query) {query.setResultTransformer(CriteriaSpecification.DISTINCT_ROOT_ENTITY);return query;}/** * 为Criteria添加distinct transformer. * 预加载关联对象的HQL会引起主对象重复, 需要进行distinct处理. */public Criteria distinct(Criteria criteria) {criteria.setResultTransformer(CriteriaSpecification.DISTINCT_ROOT_ENTITY);return criteria;}/** * 取得对象的主键名. */public String getIdName() {ClassMetadata meta = getSessionFactory().getClassMetadata(entityClass);return meta.getIdentifierPropertyName();}/** * 判断对象的属性值在数据库内是否唯一. *  * 在修改对象的情景下,如果属性新修改的值(value)等于属性原来的值(orgValue)则不作比较. */public boolean isPropertyUnique(final String propertyName, final Object newValue, final Object oldValue) {if (newValue == null || newValue.equals(oldValue)) {return true;}Object object = findUniqueBy(propertyName, newValue);return (object == null);}@Overridepublic void update(T object) {getSession().update(object);}@SuppressWarnings("unchecked")@Overridepublic T load(ID id) {return (T) getSession().load(this.entityClass, id);}/** * 将list转化为数组 * @param list * @return */protected Criterion[] list2Array(List<Criterion> list){if(list == null){return new Criterion[0];}Criterion[] result = new Criterion[list.size()];for(int i = 0; i < list.size(); i++){result[i] = list.get(i);}return result;}@Overridepublic void afterPropertiesSet() throws Exception {templateCache = new HashMap<String, StatementTemplate>();if(this.dynamicStatementBuilder == null){this.dynamicStatementBuilder = new NoneDynamicHibernateStatementBuilder();}dynamicStatementBuilder.init();Map<String,String> namedHQLQueries = dynamicStatementBuilder.getNamedHQLQueries();Map<String,String> namedSQLQueries = dynamicStatementBuilder.getNamedSQLQueries();Configuration configuration = new Configuration();configuration.setNumberFormat("#");StringTemplateLoader stringLoader = new StringTemplateLoader();for(Entry<String, String> entry : namedHQLQueries.entrySet()){stringLoader.putTemplate(entry.getKey(), entry.getValue());templateCache.put(entry.getKey(), new StatementTemplate(StatementTemplate.TYPE.HQL,new Template(entry.getKey(),new StringReader(entry.getValue()),configuration)));}for(Entry<String, String> entry : namedSQLQueries.entrySet()){stringLoader.putTemplate(entry.getKey(), entry.getValue());templateCache.put(entry.getKey(), new StatementTemplate(StatementTemplate.TYPE.SQL,new Template(entry.getKey(),new StringReader(entry.getValue()),configuration)));}configuration.setTemplateLoader(stringLoader);}protected String processTemplate(StatementTemplate statementTemplate,Map<String, ?> parameters){StringWriter stringWriter = new StringWriter();try {statementTemplate.getTemplate().process(parameters, stringWriter);} catch (Exception e) {LOGER.error("处理DAO查询参数模板时发生错误：{}",e.toString());throw new SysException(e);}return stringWriter.toString();}}

我们的SimpleHibernateDAO实现了InitializingBean,在其afterProperties方法中我们将调用DynamicHibernateStatementBuilder把语句缓存起来
上层方法调用示例-这个已经非常类似ibatis了

public List<Resource> getDescendants(Long userId,String code) {Map<String, Object> values = new HashMap<String, Object>();values.put("userId", String.valueOf(userId));values.put("code", code);values.put("type", String.valueOf(ResourceTypeEnum.URL_RESOURCE.getType()));values.put("status", String.valueOf(ResourceStatusEnum.ACTIVE.getStatus()));return this.findByNamedQuery(new ResourceRowMapper(),"resource.getDescendants", values);}