2019独角兽企业重金招聘Python工程师标准>>>
Spring提供了丰富的标签和注解来进行bean的定义,除此之外框架来提供了扩展机制让使用可以通过properties来定义bean,与强大的标签式和注解式的bean定义相比,properties提供的规则要简单许多。
key部分用.分隔即通过A.B来进行相关的属性定义,其中A表示bean名称,B通过不同的值还表达不同的含义:
- (class),bean的类型
- (parent),bean的父bean
- name,bean的name属性,name是一个普通属性
- childBean(ref),bean的childBean属性,childBean是一个引用属性
- (singleton),是否单例
- (lazy-init),是否懒加载
- $0,第一个构造子参数
- (scope),作用域
- (abstract),是否是抽象bean
看一个例子:
bean.properties文件
[java] view plain copy
- #bean1
- propBean.(class) = spring.beans.properties.PropBean
- propBean.(parent) = commonBean
- propBean.name = name1
- propBean.childBean(ref) = childBean
- propBean.(singleton) = true
- propBean.(lazy-init) = true
- #bean2
- childBean.(class) = spring.beans.properties.ChildBean
- childBean.$0 = cid1
- chlldBean.(scope) = singleton
- #abstract bean
- commonBean.(class) = spring.beans.properties.CommonBean
- commonBean.id = 1
- commonBean.(abstract) = true
上面的properties文件定义了三个bean:
- commonBean,类型是spring.beans.properties.CommonBean,注入值1到id属性,这是一个抽象bean
- childBean,类型spring.beans.properties.ChildBean,构造器注入cid1,作用域是singleton
- propBean,类型是spring.beans.properties.PropBean,父bean是commonBean,注入一个普通属性name,和引用属性childBean,引用的bean是childBean,bean是单例并且懒加载。
bean定义文件写好之后,通过PropertiesBeanDefinitionReader来加载解析bean定义,这个解析器的原理很简单,在此不做详细分析,下面是实例代码。
[java] view plain copy
- public void test() {
- GenericApplicationContext ctx = new GenericApplicationContext();
- Resource res = new ClassPathResource(
- "spring/beans/properties/bean.properties");
- PropertiesBeanDefinitionReader propReader = new PropertiesBeanDefinitionReader(
- ctx);
- propReader.loadBeanDefinitions(res);
- PropBean propBean = (PropBean) ctx.getBean("propBean");
- assertNotNull(propBean);
- assertNotNull(propBean.getId());
- assertNotNull(propBean.getChildBean());
- assertNotNull(propBean.getChildBean().getCid());
- }
也可以完全不用单独定义个properties文件,只需要把相关的key-value放到一个Map中,再通过PropertiesBeanDefinitionReader加载这个map中的key-value。
通过这种方式,使用者可以根据需要自定义些bean的定义规则,比如可以把bean定义放在数据库中,把数据库中的信息读取出来拼接成满足properties规则的bean定义,在Spring中就定义了一个org.springframework.jdbc.core.support.JdbcBeanDefinitionReader来完成这种需求,看下这个类的代码。
[java] view plain copy
- public class JdbcBeanDefinitionReader {
- private final PropertiesBeanDefinitionReader propReader;
- private JdbcTemplate jdbcTemplate;
- /**
- * Create a new JdbcBeanDefinitionReader for the given bean factory,
- * using a default PropertiesBeanDefinitionReader underneath.
- * <p>DataSource or JdbcTemplate still need to be set.
- * @see #setDataSource
- * @see #setJdbcTemplate
- */
- public JdbcBeanDefinitionReader(BeanDefinitionRegistry beanFactory) {
- this.propReader = new PropertiesBeanDefinitionReader(beanFactory);
- }
- /**
- * Create a new JdbcBeanDefinitionReader that delegates to the
- * given PropertiesBeanDefinitionReader underneath.
- * <p>DataSource or JdbcTemplate still need to be set.
- * @see #setDataSource
- * @see #setJdbcTemplate
- */
- public JdbcBeanDefinitionReader(PropertiesBeanDefinitionReader beanDefinitionReader) {
- Assert.notNull(beanDefinitionReader, "Bean definition reader must not be null");
- this.propReader = beanDefinitionReader;
- }
- /**
- * Set the DataSource to use to obtain database connections.
- * Will implicitly create a new JdbcTemplate with the given DataSource.
- */
- public void setDataSource(DataSource dataSource) {
- this.jdbcTemplate = new JdbcTemplate(dataSource);
- }
- /**
- * Set the JdbcTemplate to be used by this bean factory.
- * Contains settings for DataSource, SQLExceptionTranslator, NativeJdbcExtractor, etc.
- */
- public void setJdbcTemplate(JdbcTemplate jdbcTemplate) {
- Assert.notNull(jdbcTemplate, "JdbcTemplate must not be null");
- this.jdbcTemplate = jdbcTemplate;
- }
- /**
- * Load bean definitions from the database via the given SQL string.
- * @param sql SQL query to use for loading bean definitions.
- * The first three columns must be bean name, property name and value.
- * Any join and any other columns are permitted: e.g.
- * {@code SELECT BEAN_NAME, PROPERTY, VALUE FROM CONFIG WHERE CONFIG.APP_ID = 1}
- * It's also possible to perform a join. Column names are not significant --
- * only the ordering of these first three columns.
- */
- public void loadBeanDefinitions(String sql) {
- Assert.notNull(this.jdbcTemplate, "Not fully configured - specify DataSource or JdbcTemplate");
- final Properties props = new Properties();
- this.jdbcTemplate.query(sql, new RowCallbackHandler() {
- public void processRow(ResultSet rs) throws SQLException {
- String beanName = rs.getString(1);
- String property = rs.getString(2);
- String value = rs.getString(3);
- // Make a properties entry by combining bean name and property.
- props.setProperty(beanName + "." + property, value);
- }
- });
- this.propReader.registerBeanDefinitions(props);
- }
- }
此外,通过理解PropertiesBeanDefinitionReader的实现方式,发现也可以通过扩展BeanDefinitionReader来扩展bean定义,我们可以通过继承AbstractBeanDefinitionReader来完成这种扩展。