ibatis 3.0 Dynamic Sql 设计解析(并与2.x的差异)

jelly_x

浏览: 162300 次
性别:
来自: 成都

最近访客更多访客>>

limingnihao

冷月宫主

jklwan

lsj626472785

博主相关

博客

微博

相册

留言

关于我

文章分类

社区版块

存档分类

博客分类：

IBatis

SQL iBATIS 设计模式 XML Blog

前段时间ibatis3.0发布出来了，迫不及待，将其源码下载拜读。相对ibatis 2.x来说，3.0已是完全改变。具体我就不在这细说，论坛中有一个帖子介绍了ibatis 3.0的新特征及使用。

由于其他模块的源码我还未细读，在这篇中，先来讨论Dynamic Sql在ibatis 3.0中的实现并比较2.x对应模块的设计。

写在前头的话：

其实如从设计模式应用角度去看待ibatis 3.0中Dynamic Sql的实现，这篇跟我的上篇(HtmlParser设计解析(1)-解析器模式 )相同，都是使用Interpreter模式。

这篇权当Interpreter模式的另一个demo，认我们体会这些开源项目中设计模式的使用。学习都是从模仿开始的，让我们吸收高人们的经验，应用于我们实践项目需求中。

从总结中提高：

一、对比2.x中与3.0的Sqlmap中dynamic sql配置

2.x:

Xml代码

< select id = "dynamicGetAccountList" parameterClass = "Account" resultClass = "Account" >
select ACC_ID as id,
ACC_FIRST_NAME as firstName,
ACC_LAST_NAME as lastName,
ACC_EMAIL as emailAddress from ACCOUNT
< dynamic prepend = "WHERE" >
< isNotNull prepend = "AND" property = "emailAddress" >
ACC_EMAIL = #emailAddress#
</ isNotNull >
< isNotNull property = "idList" prepend = " or ACC_ID in " >
< iterate property = "idList" conjunction = "," open = "(" close = ")" >
#id#
</ iterate >
</ isNotNull >
</ dynamic >
</ select >

<select id="dynamicGetAccountList" parameterClass="Account" resultClass="Account"> 
      select ACC_ID as id,
      ACC_FIRST_NAME as firstName,
      ACC_LAST_NAME as lastName,
      ACC_EMAIL as emailAddress from ACCOUNT
	  
	<dynamic prepend="WHERE">
	  <isNotNull prepend="AND" property="emailAddress">
        ACC_EMAIL = #emailAddress#
      </isNotNull>
      <isNotNull property="idList" prepend=" or ACC_ID in ">   
		<iterate property="idList" conjunction="," open="(" close=")" >   
			#id#  
		</iterate>   
	  </isNotNull>   
    </dynamic>
</select>

3.0:

Xml代码

< select id = "dynamicGetAccountList" parameterType = "Account" resultType = "Account" >
select ACC_ID as id,
ACC_FIRST_NAME as firstName,
ACC_LAST_NAME as lastName,
ACC_EMAIL as emailAddress from ACCOUNT
< where >
< if test = "emailAddress != null" > ACC_EMAIL = #{emailAddress} </ if >
< if test = "idList != null" >
or ACC_ID IN
< foreach item = "id" index = "index" open = "(" close = ")" separator = "," collection = "idList" >
#{idList[${index}]}
</ foreach >
</ if >
</ where >
</ select >

<select id="dynamicGetAccountList" parameterType="Account" resultType="Account">
	  select ACC_ID as id,
	  ACC_FIRST_NAME as firstName,
	  ACC_LAST_NAME as lastName,
	  ACC_EMAIL as emailAddress from ACCOUNT

	<where>
		<if test="emailAddress != null">ACC_EMAIL = #{emailAddress}</if>
		<if test="idList != null">
			or ACC_ID IN
			<foreach item="id" index="index" open="(" close=")" separator="," collection="idList">
				#{idList[${index}]}
			</foreach>
	   </if>
	</where>
</select>

从上面这个简单的比较中，第一感觉3.0了中其dynamic sql更加简洁明了。

其二，test="emailAddress != null" 添加了OGNL的解释支持，可以动态支持更多的判断，这将不限于原2.x中提供

的判断逻辑，更不需要为每个判断条件加个标签进行配置。

例如：<if test="id > 10 && id < 20"> ACC_EMAIL = #{emailAddress}</if>

<if test="Account.emailAddress != null "> ACC_EMAIL = #{emailAddress}</if> ……

二、2.x Dynamic Sql的设计

2.1、2.x中dynamic流程。

这里帖出，我先前在分析ibatis 2.3时画的一个对dynamic sql的整体使用的时序图，可能会显得乱而复杂。

2.2、主要类设计

在这，我们只关注这几个类：XMLSqlSource、DynamicSql、SqlTagHandler (具体类结构图见后)

XMLSqlSource：相当于一个工厂类，其核心方法parseDynamicTags()，用于解析sql Tag，并判断是否是动态SQL标签。如果true，返回一个DynamicSql对象并创建多个SqlChildt对象添加至动态SQL列表中(addChild())；false，返回RawSql对象(简单的SQL语句) 。

DynamicSql：核心的动态SQL类。其动态条件判断逻辑，参数映射等都发生在这个类中。

SqlTagHandle：动态条件判断接口，其每个动态SQL标签对应其一个子类。

接下来，我们具体看下在DynamicSql类中核心方法。

DynamicSql:

Java代码

private void processBodyChildren(StatementScope statementScope, SqlTagContext ctx, Object parameterObject, Iterator localChildren, PrintWriter out) {
while (localChildren.hasNext()) { //XMLSqlSource 生成的动态SQL列表
SqlChild child = (SqlChild) localChildren.next();
if (child instanceof SqlText) {
... ... //组装SQL语句及映射SQL参数
} else if (child instanceof SqlTag) {
SqlTag tag = (SqlTag) child;
SqlTagHandler handler = tag.getHandler(); //得到动态SQL标签处理器
int response = SqlTagHandler.INCLUDE_BODY;
do {
response = handler.doStartFragment(ctx, tag, parameterObject); //处理开始片段
if (response != SqlTagHandler.SKIP_BODY) { //是否跳过，意思该判断的条件为false
processBodyChildren(statementScope, ctx, parameterObject, tag.getChildren(), pw); //递归处理
StringBuffer body = sw.getBuffer();
response = handler.doEndFragment(ctx, tag, parameterObject, body); //处理结束片段
handler.doPrepend(ctx, tag, parameterObject, body); //组装SQL
}
} while (response == SqlTagHandler.REPEAT_BODY);
... ... }
}

private void processBodyChildren(StatementScope statementScope, SqlTagContext ctx, Object parameterObject, Iterator localChildren, PrintWriter out) {
    while (localChildren.hasNext()) { //XMLSqlSource 生成的动态SQL列表
      SqlChild child = (SqlChild) localChildren.next();
      if (child instanceof SqlText) {
          ... ... //组装SQL语句及映射SQL参数
      } else if (child instanceof SqlTag) {
        SqlTag tag = (SqlTag) child;
        SqlTagHandler handler = tag.getHandler(); //得到动态SQL标签处理器
        int response = SqlTagHandler.INCLUDE_BODY;
        do {          
          response = handler.doStartFragment(ctx, tag, parameterObject); //处理开始片段
          if (response != SqlTagHandler.SKIP_BODY) { //是否跳过，意思该判断的条件为false
            processBodyChildren(statementScope, ctx, parameterObject, tag.getChildren(), pw); //递归处理
            StringBuffer body = sw.getBuffer();
            response = handler.doEndFragment(ctx, tag, parameterObject, body); //处理结束片段
            handler.doPrepend(ctx, tag, parameterObject, body); //组装SQL
            
          }
        } while (response == SqlTagHandler.REPEAT_BODY);
		... ...    }
}

2.3、SqlTagHandle设计

首先看下SqlTagHandle处理类的结果图：

ConditionalTagHandler：

Java代码

public abstract class ConditionalTagHandler extends BaseTagHandler {
... ...
public abstract boolean isCondition(SqlTagContext ctx, SqlTag tag, Object parameterObject);
public int doStartFragment(SqlTagContext ctx, SqlTag tag, Object parameterObject) {
ctx.pushRemoveFirstPrependMarker(tag);
if (isCondition(ctx, tag, parameterObject)) {
return SqlTagHandler.INCLUDE_BODY;
} else {
return SqlTagHandler.SKIP_BODY;
}
}
... ...
}

public abstract class ConditionalTagHandler extends BaseTagHandler {
  ... ...
  public abstract boolean isCondition(SqlTagContext ctx, SqlTag tag, Object parameterObject);

  public int doStartFragment(SqlTagContext ctx, SqlTag tag, Object parameterObject) {
    ctx.pushRemoveFirstPrependMarker(tag);
    if (isCondition(ctx, tag, parameterObject)) {
      return SqlTagHandler.INCLUDE_BODY;
    } else {
      return SqlTagHandler.SKIP_BODY;
    }
  }
  ... ...
}

IsNullTagHandler：

Java代码

public class IsNullTagHandler extends ConditionalTagHandler {
private static final Probe PROBE = ProbeFactory.getProbe();
public boolean isCondition(SqlTagContext ctx, SqlTag tag, Object parameterObject) {
if (parameterObject == null ) {
return true ;
} else {
String prop = getResolvedProperty(ctx, tag);
Object value;
if (prop != null ) {
value = PROBE.getObject(parameterObject, prop);
} else {
value = parameterObject;
}
return value == null ;
}
}
}

public class IsNullTagHandler extends ConditionalTagHandler {
	private static final Probe PROBE = ProbeFactory.getProbe();
	public boolean isCondition(SqlTagContext ctx, SqlTag tag, Object parameterObject) {
		if (parameterObject == null) {
			return true;
		} else {
			String prop = getResolvedProperty(ctx, tag);
			Object value;
			if (prop != null) {
				value = PROBE.getObject(parameterObject, prop);
			} else {
				value = parameterObject;
			}
			return value == null;
		}
	}
}

至于其他的相关类，不在这列出了，有兴趣的可以找其源码了解下。

2.4、总结ibatis 2.X Dynamic Sql 的设计

从上面的分析中，可以体会出作者的dynamic sql这模块的设计思路。从装载sqlmap.xml中各sql配置(时序图中的1步)，通过工厂创建DynamicSql和RawSql(时序图中的3步)，然后分发之不同的处理器。

在DynamicSql中则调用SqlTagHandle判断其条件(时序图中的10步)。而SqlTagHandle的设计使用策略者模式，让其不同的子类来处理这个判断逻辑。

通过一系列的加工，最终组装一个Sql对象，将值set至MappedStatement(时序图中的14步)中，然后MappedStatement对象执行executeQueryWithCallback查询数据(时序图中的17步)，这儿会调用先前组装的Sql对象(时序图中的19步)。至于这其中的细节已不在这篇的研究这内。

三、3.0 Dynamic Sql的设计

至于3.0其基本流程跟2.x是一样的，从装载 -> 参数映射 -> 执行SQL -> 返回结果。我们直接切入主题，分析是核心部分。先从一个简单的Dynamic Sql的测试用例开始。

3.1、测试用例

dynamic sql test:

Java代码

@Test
public void shouldTrimWHEREInsteadOfORForSecondCondition() throws Exception {
/* SELECT * FROM BLOG
<where>
<if test="id != false"> and ID = #{id} </if>
<if test="name != false"> or NAME = #{name} </if>
</where>
*/
final String expected = "SELECT * FROM BLOG WHERE NAME = ?" ;
DynamicSqlSource source = createDynamicSqlSource(
new TextSqlNode( "SELECT * FROM BLOG" ),
new WhereSqlNode(mixedContents(
new IfSqlNode(
mixedContents( new TextSqlNode( " and ID = ? " )), "false" ), new IfSqlNode(mixedContents( new TextSqlNode( " or NAME = ? " )), "true" ))));
BoundSql boundSql = source.getBoundSql( null );
assertEquals(expected, boundSql.getSql());
}
private DynamicSqlSource createDynamicSqlSource(SqlNode... contents)
throws IOException, SQLException {
createBlogDataSource();
final String resource = ".../MapperConfig.xml" ;
final Reader reader = Resources.getResourceAsReader(resource);
SqlSessionFactory sqlMapper = new SqlSessionFactoryBuilder()
.build(reader);
Configuration configuration = sqlMapper.getConfiguration();
MixedSqlNode sqlNode = mixedContents(contents);
return new DynamicSqlSource(configuration, sqlNode);
}
private MixedSqlNode mixedContents(SqlNode... contents) {
return new MixedSqlNode(Arrays.asList(contents));
}

	@Test
	public void shouldTrimWHEREInsteadOfORForSecondCondition() throws Exception {
	 /*	SELECT * FROM BLOG
		<where>
			<if test="id != false"> and ID = #{id} </if>
			<if test="name != false"> or NAME = #{name} </if>
		</where>
	   */
		final String expected = "SELECT * FROM BLOG WHERE  NAME = ?";
		DynamicSqlSource source = createDynamicSqlSource(
				new TextSqlNode("SELECT * FROM BLOG"), 
					new WhereSqlNode(mixedContents(
							new IfSqlNode(
									mixedContents(new TextSqlNode("   and ID = ?  ")),"false"), new IfSqlNode(mixedContents(new TextSqlNode("   or NAME = ?  ")), "true"))));
		BoundSql boundSql = source.getBoundSql(null);
		assertEquals(expected, boundSql.getSql());
	}

	private DynamicSqlSource createDynamicSqlSource(SqlNode... contents)
			throws IOException, SQLException {
		createBlogDataSource();
		final String resource =  ".../MapperConfig.xml";
		final Reader reader = Resources.getResourceAsReader(resource);
		SqlSessionFactory sqlMapper = new SqlSessionFactoryBuilder()
				.build(reader);
		Configuration configuration = sqlMapper.getConfiguration();
		MixedSqlNode sqlNode = mixedContents(contents);
		return new DynamicSqlSource(configuration, sqlNode);
	}

	private MixedSqlNode mixedContents(SqlNode... contents) {
		return new MixedSqlNode(Arrays.asList(contents));
	}

有经验的人，我想一眼就能看出其3.0中的设计思想，从Test中可以看出，或者我上一篇介绍的HtmlParser NodeFilter。

YES，在ibatis 3.0 dynamic sql设计正是应用了解释器模式，替换了原在这种需求下相对显得笨拙的策略者模式。

下面具体看下类结构图。

3.2、类结构图

SqlNode Class Diagram :

SqlSource Class Diagram :

3.3、配置文件的解析

在这，我就顺便提下ibatis解析组件对dynamic sql的解析方式，以代码见分晓吧。

XMLStatementBuilder：

Java代码

public void parseStatementNode(XNode context) {
... ...
List<SqlNode> contents = parseDynamicTags(context);
MixedSqlNode rootSqlNode = new MixedSqlNode(contents); //再次包装dynamic sql处理链
SqlSource sqlSource = new DynamicSqlSource(configuration, rootSqlNode); //默认初始化DynamicSqlSource
... ...
builderAssistant.addMappedStatement(id, sqlSource, statementType,
sqlCommandType, fetchSize, timeout, parameterMap,
parameterTypeClass, resultMap, resultTypeClass,
resultSetTypeEnum, flushCache, useCache, keyGenerator,
keyProperty); //将解析的所有属性构建成相应的对象存入全局的申明对象(MappedStatement)中，后面只传递该对象。
}
private List<SqlNode> parseDynamicTags(XNode node) {
List<SqlNode> contents = new ArrayList<SqlNode>();
NodeList children = node.getNode().getChildNodes();
for ( int i = 0 ; i < children.getLength(); i++) {
XNode child = node.newXNode(children.item(i));
String nodeName = child.getNode().getNodeName();
if (child.getNode().getNodeType() == Node.CDATA_SECTION_NODE
|| child.getNode().getNodeType() == Node.TEXT_NODE) {
String data = child.getStringBody( "" );
contents.add( new TextSqlNode(data));
} else {
NodeHandler handler = nodeHandlers.get(nodeName);
if (handler == null ) {
throw new BuilderException( "Unknown element <" + nodeName "> in SQL statement." );
}
handler.handleNode(child, contents);
}
}
return contents;
}
private Map<String, NodeHandler> nodeHandlers = new HashMap<String, NodeHandler>() {
{
put( "where" , new WhereHandler());
put( "set" , new SetHandler());
put( "foreach" , new ForEachHandler());
put( "if" , new IfHandler());
... ...
}
};
private interface NodeHandler {
void handleNode(XNode nodeToHandle, List<SqlNode> targetContents);
}
private class WhereHandler implements NodeHandler {
public void handleNode(XNode nodeToHandle, List<SqlNode> targetContents) {
List<SqlNode> contents = parseDynamicTags(nodeToHandle); // 遍历
MixedSqlNode mixedSqlNode = new MixedSqlNode(contents); //对应测试用例中的mixedContents方法
WhereSqlNode where = new WhereSqlNode(mixedSqlNode);
targetContents.add(where);
}
}
private class IfHandler implements NodeHandler {
public void handleNode(XNode nodeToHandle, List<SqlNode> targetContents) {
List<SqlNode> contents = parseDynamicTags(nodeToHandle); //遍历
MixedSqlNode mixedSqlNode = new MixedSqlNode(contents);
String test = nodeToHandle.getStringAttribute( "test" );
IfSqlNode ifSqlNode = new IfSqlNode(mixedSqlNode, test); //初始化对应的处理器
targetContents.add(ifSqlNode); //
}
} // 其他的Handle详见ibatis源码~

	public void parseStatementNode(XNode context) {
                ...  ...
		List<SqlNode> contents = parseDynamicTags(context);
		MixedSqlNode rootSqlNode = new MixedSqlNode(contents);//再次包装dynamic sql处理链
		SqlSource sqlSource = new DynamicSqlSource(configuration, rootSqlNode); //默认初始化DynamicSqlSource
                ... ...
		builderAssistant.addMappedStatement(id, sqlSource, statementType,
				sqlCommandType, fetchSize, timeout, parameterMap,
				parameterTypeClass, resultMap, resultTypeClass,
				resultSetTypeEnum, flushCache, useCache, keyGenerator,
				keyProperty); //将解析的所有属性构建成相应的对象存入全局的申明对象(MappedStatement)中，后面只传递该对象。
	}

	private List<SqlNode> parseDynamicTags(XNode node) {
		List<SqlNode> contents = new ArrayList<SqlNode>();
		NodeList children = node.getNode().getChildNodes();
		for (int i = 0; i < children.getLength(); i++) {
			XNode child = node.newXNode(children.item(i));
			String nodeName = child.getNode().getNodeName();
			if (child.getNode().getNodeType() == Node.CDATA_SECTION_NODE
					|| child.getNode().getNodeType() == Node.TEXT_NODE) {
				String data = child.getStringBody("");
				contents.add(new TextSqlNode(data));
			} else {
				NodeHandler handler = nodeHandlers.get(nodeName);
				if (handler == null) {
					throw new BuilderException("Unknown element <" + nodeName "> in SQL statement.");
				}
				handler.handleNode(child, contents);
			}
		}
		return contents;
	}         
	private Map<String, NodeHandler> nodeHandlers = new HashMap<String, NodeHandler>() {
		{
			put("where", new WhereHandler());
			put("set", new SetHandler());
			put("foreach", new ForEachHandler());
			put("if", new IfHandler());
			... ...
		}
	};
	private interface NodeHandler {
		void handleNode(XNode nodeToHandle, List<SqlNode> targetContents);
	}
	private class WhereHandler implements NodeHandler {
		public void handleNode(XNode nodeToHandle, List<SqlNode> targetContents) {
			List<SqlNode> contents = parseDynamicTags(nodeToHandle);// 遍历
			MixedSqlNode mixedSqlNode = new MixedSqlNode(contents);//对应测试用例中的mixedContents方法
			WhereSqlNode where = new WhereSqlNode(mixedSqlNode);
			targetContents.add(where);
		}
	}
	private class IfHandler implements NodeHandler {
		public void handleNode(XNode nodeToHandle, List<SqlNode> targetContents) {
			List<SqlNode> contents = parseDynamicTags(nodeToHandle);//遍历
			MixedSqlNode mixedSqlNode = new MixedSqlNode(contents);
			String test = nodeToHandle.getStringAttribute("test");
			IfSqlNode ifSqlNode = new IfSqlNode(mixedSqlNode, test);//初始化对应的处理器
			targetContents.add(ifSqlNode);//
		}
	} // 其他的Handle详见ibatis源码~

上面是其解析代码的一部分，我想从这几行代码中，可以看出作者的思想了(遍历XML各节点，以节点名查找相应对应的处理器，分发之该处理器执行"业务分析" — 策略者模式，这样在XML中定义了多少标签，这里就需要多少个类与之对应，但如果策略类太多，这种方式就显得笨拙了)。

以下就是其核心类的一部分源码，先看再说。

3.4、DynamicSqlSource(核心类)

Java代码

public class DynamicSqlSource implements SqlSource {
public DynamicSqlSource(Configuration configuration, SqlNode rootSqlNode) {
this .configuration = configuration;
this .rootSqlNode = rootSqlNode;
}
public BoundSql getBoundSql(Object parameterObject) {
DynamicContext context = new DynamicContext(parameterObject); //组装后的结果存储类
rootSqlNode.apply(context); //调用SqlNode解释sql,并组装成完整的sql(SqlNode的客户端调用就在这)
SqlSourceBuilder sqlSourceParser = new SqlSourceBuilder(configuration);
Class parameterType = parameterObject == null ? Object. class : parameterObject.getClass();
SqlSource sqlSource = sqlSourceParser.parse(context.getSql(), parameterType);
BoundSql boundSql = sqlSource.getBoundSql(parameterObject);
for (Map.Entry<String, Object> entry : context.getBindings().entrySet()) {
boundSql.setAdditionalParameter(entry.getKey(), entry.getValue());
}
return boundSql;
}
}

public class DynamicSqlSource implements SqlSource {
	public DynamicSqlSource(Configuration configuration, SqlNode rootSqlNode) {
		this.configuration = configuration;
		this.rootSqlNode = rootSqlNode;
	}
	public BoundSql getBoundSql(Object parameterObject) {
		DynamicContext context = new DynamicContext(parameterObject);//组装后的结果存储类
		rootSqlNode.apply(context);//调用SqlNode解释sql,并组装成完整的sql(SqlNode的客户端调用就在这)
		SqlSourceBuilder sqlSourceParser = new SqlSourceBuilder(configuration);
		Class parameterType = parameterObject == null ? Object.class : parameterObject.getClass();
		SqlSource sqlSource = sqlSourceParser.parse(context.getSql(), parameterType);
		BoundSql boundSql = sqlSource.getBoundSql(parameterObject);
		for (Map.Entry<String, Object> entry : context.getBindings().entrySet()) {
			boundSql.setAdditionalParameter(entry.getKey(), entry.getValue());
		}
		return boundSql;
	}
}

3.5、SqlNode

Java代码

public interface SqlNode {
public boolean apply(DynamicContext context);
}

public interface SqlNode {
	public boolean apply(DynamicContext context);
}

MixedSqlNode.class

Java代码

public class MixedSqlNode implements SqlNode {
... ....
public boolean apply(DynamicContext context) {
//遍历组装的解析内容
for (SqlNode sqlNode : contents) {
// 转发至相关解释器处理
sqlNode.apply(context);
}
return true ;
}
}

public class MixedSqlNode implements SqlNode {
        ... ....
	public boolean apply(DynamicContext context) {
		//遍历组装的解析内容
		for (SqlNode sqlNode : contents) { 
			// 转发至相关解释器处理
			sqlNode.apply(context); 
		}
		return true;
	}
}

IfSqlNode.class

Java代码 <

分享到：

StringUtils工具类的使用 | Java获得系统当前年份日期

2011-03-18 21:42
浏览 1096
评论(0)
分类:企业架构
查看更多

发表评论

您还没有登录,请您登录后再发表评论

最近访客更多访客>>

博主相关

文章分类

社区版块

存档分类

最新评论