为了避免sql编写在Java代码里面,所以实现类型mybaits的功能,把sql语句编写在xml文件中,这样可以统一管理sql语句,维护更加容易。
1. 首先编写配置sql语句的xml的dtd文件,dtd文件主要是规范xml的,在编写sql语句的配置文件中主要有五个标签:select , update , insert , delete , import
其中select有两个属性id(保证所有的sql语句id唯一),resultClass(查询语句返回的对象,可以使具体的实体类,也可以是Map,List);import只有一个resource属性,指定导入的xml的位置,其他的标签都只有一个id属性。dtd文件如下,存放在hqhop-framework-common项目的/src/main/resources/META-INF下:
</pre><pre name="code" class="html"><!ELEMENT sqls-configs (import | insert | delete | update | select )*> <!ELEMENT insert (#PCDATA)> <!ATTLIST insert id ID #REQUIRED> <!ELEMENT delete (#PCDATA)> <!ATTLIST delete id ID #REQUIRED> <!ELEMENT update (#PCDATA)> <!ATTLIST update id ID #REQUIRED> <!ELEMENT select (#PCDATA)> <!ATTLIST select resultClass CDATA #REQUIRED> <!ATTLIST select id ID #REQUIRED> <!ELEMENT import EMPTY> <!ATTLIST import resource CDATA #REQUIRED>
2. sql配置文件的参考事例:
<?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE sqls-configs PUBLIC "-//SQLS//MAIN CONFIG DTD//EN" "http://kf.hqhop.com:8888/hqhop-framework-sqls.dtd"> <sqls-configs> <select <span style="color:#ff0000;">id="select1"</span> resultClass="<span style="color:#ff0000;">com.hqhop.framework.shiro.entity.User</span>"> select * from t_base_user where id =:id </select><pre name="code" class="html"><span style="white-space:pre"> </span><select <span style="color:#ff0000;">id="select2"</span> resultClass="<span style="color:#ff0000;">java.util.Map</span>"> select * from t_base_user where id = ? </select>
<pre name="code" class="html"><span style="white-space:pre"> </span><select <span style="color:#ff0000;">id="select3" </span>resultClass="<span style="color:#ff0000;">java.util.List</span>"> select * from t_base_user where id = ? </select>
<update id="update"></update><insert id="insert"></insert><delete id="delete"></delete><import resource="classpath:sql/test.xml" /></sqls-configs>
3. 解析sql的配置文件,类目录结构
实现主要就用到了这4个类,首先把每一条sql语句的信息存放到SQL对象中,然后在把SQL对象和配置的sql的id对应存放到SQLContext,当我们在使用的时候就直接去SQLContext中去取。
SQL.java
public class SQL { private String id; //sql的id private String sql; //具体的sql private Class resultClass; //返回的实体类,只对select标签有效 public SQL(String id, String sql, String resultClass) { this.id = id; this.sql = sql; try { if (Utils.isNotEmpty(resultClass)) this.resultClass = Class.forName(resultClass); else this.resultClass = null; } catch (ClassNotFoundException e) { e.printStackTrace(); } }<pre name="code" class="java"> //get set....
}
SQLContext.java
public class SQLContext { <span style="white-space:pre"> </span>//仅仅一个属性,使用map存放每一个sql语句,使用的时候更加key来找到sql private Map<String, SQL> sqls = new HashMap<>(); public SQL getSQL(String sqlKey) { return sqls.get(sqlKey); } public Set<String> getKeys() { return sqls.keySet(); } public void put(String sqlKey, SQL sql) { this.sqls.put(sqlKey, sql); } public boolean checkKeyIsExist(String key) { return this.sqls.containsKey(key); } }
4. 解析sql的xml文件主要是在SQLContextFileBuilder中完成的,是一个的单例模式
1) 首先解析import标签,使用递归的方式把所有import的xml加载到一个document,实现的方法是public Document loadFullConfigFile(Resource resource, String encoding)
// 解析import包含的子配置文件 public Document loadFullConfigFile(Resource resource, String encoding) throws UnsupportedEncodingException, IOException, DocumentException { SAXReader reader = null; Document document = null; reader = new SAXReader(); InputStreamReader isr = null; try { isr = new InputStreamReader(resource.getInputStream(), encoding); document = reader.read(isr); } finally { CloseUtil.close(isr); } final Element root = document.getRootElement(); List list = document.selectNodes("//import"); for (int i = 0; i < list.size(); i++) { Element n = (Element) list.get(i); String file = n.attribute("resource").getValue(); Resource fr = SpringUtils.getResource(file); <span style="color:#ff0000;">Document includedDoc = loadFullConfigFile(fr, encoding);</span> List content = root.content(); int indexOfPos = content.indexOf(n); content.remove(indexOfPos); Element ie = includedDoc.getRootElement(); List ie_children = ie.content(); for (int k = ie_children.size() - 1; k >= 0; k--) { content.add(indexOfPos, ie_children.get(k)); } } this.rootDoc = document.getRootElement(); return document; }
2) 再从document中解析出所有的sql语句,创建SQL对象,存放到SQLContext中
public SQLContext loadSQLContext() { SQLContext sqlContext = new SQLContext(); List els = rootDoc.elements(); for (Object o : els) { Element element = (Element) o; String idKey = element.attribute("id").getValue(); String sql = element.getTextTrim(); Attribute attribute = element.attribute("resultClass"); String resultClass = Utils.isNotEmpty(attribute) ? attribute.getValue() : null; <span style="color:#ff0000;">//判断sql的id是否已经存在 ,若存在就抛出异常</span> if (<span style="color:#ff0000;">!sqlContext.checkKeyIsExist(idKey)</span>) { sqlContext.put(idKey, new SQL(idKey, sql, resultClass)); } else { throw new RuntimeException("请检查sql的配置文件,sql的key已经存在(key=" + idKey + ")!"); } } return sqlContext; }
5. 让spring容器启动就去加载sql的xml文件:SQLContextFactory.java 只有一个方法,
public static SQLContext createSQLContext(Resource springResource){......} 完成解析xml文件,返回sqlContext
在spring的xml配置文件中添加如下代码来完成SQLContext的创建:
<bean id="sqlContext" class="com.hqhop.framework.common.orm.sql.SQLContextFactory" factory-method="createSQLContext"> <constructor-arg> <!-- 指定sql文件的位置 --> <value><span style="color:#ff0000;">classpath:sqls/sql.xml</span></value> </constructor-arg> </bean>
6. 到此为止sqlContext已经存在了spring容器中了,接下来就在BaseRepoitoryImpl中来使用sqlContext,在BaseRepoitory中定义了几个使用sql的xml的接口,这里就只说明一个接口就行了:
public List<T> findAll(String sqlKey, Object... params) {....}
@Autowired private SQLContext sqlContext; public List<T> findAll(String sqlKey, Object... params) { SQL sql = this.sqlContext.getSQL(sqlKey); SQLQuery sqlQuery = this.getSession().createSQLQuery(sql.getSql()); RepositoryHelper.setValues(sqlQuery, params); RepositoryHelper.setResultTransformer(sqlQuery, sql.getResultClass()); return sqlQuery.list(); }
1) 传入一个sql的id,然后通过sqlContext来获取对应的SQL对象,通过RepoitoryHelper.setValues(...)来设置参数的值,这里传入的params可以使多个参数值,也可以是一个map对象,具体使用什么就要根据编写的sql来决定
select * from t_base_user where id =:id 这种方式只能传入map对象
select * from t_base_user where id = ? 这种方式只能传入对个参数值
RepositoryHelper.setValues(.....)代码如下:
public static void setValues(Query query, Object... params) { if (params[0] instanceof Map) { Map mapParams = (Map) params[0]; Set<String> keys = mapParams.keySet(); for (String key : keys) { query.setParameter(key, mapParams.get(key)); } } else { int paramIndex = 0; for (Object o : params) { query.setParameter(paramIndex++, o); } } }
2) 绑定查询出来的数据应该转化成什么对象,使用RepositoryHelper.setResultTransformer(...)来完成的,这里的resultClass是在sql的xml文件中指定的,只对select标签有效:
public static void setResultTransformer(SQLQuery query, Class resultClass) { if (Utils.isNotEmpty(resultClass)) { if (<span style="color:#ff0000;">resultClass.equals(List.class)</span>) { query.setResultTransformer(Transformers.TO_LIST); } else if (<span style="color:#ff0000;">resultClass.equals(Map.class)</span>) { query.setResultTransformer(Transformers.ALIAS_TO_ENTITY_MAP); } else { query.addEntity(resultClass); } } }
其他使用sql的xml接口的实现方式都类似。