JdbcTemplate源码解析之query操作

咱们上次对执行增删改的操作进行分析以后,这次则来对查询进行流程梳理把~
源码解析


接下来则来对JdbcTemplate的query方法进行一个源码解析吧~
eg:
    List<User> users = template.query(sql,new BeanPropertyRowMapper<User>(User.class),params.toArray())
这里我们来分析下为什么返回值是List<User>
第一步是创建JdbcTemplate对象,传入数据源,进入构造方法查看

JdbcTemplate.java

[Java] 纯文本查看 复制代码

?

1 2 3 4

public JdbcTemplate(DataSource dataSource) {     setDataSource(dataSource);     afterPropertiesSet(); }

JdbcAccessor.java

[Java] 纯文本查看 复制代码

?

01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19

private DataSource dataSource;  private boolean lazyInit = true;  public void setDataSource(DataSource dataSource) {      this.dataSource = dataSource;  }  public DataSource getDataSource() {      return this.dataSource;  }  public boolean isLazyInit() {      return this.lazyInit;  }  public void afterPropertiesSet() {      if (getDataSource() == null) {          throw new IllegalArgumentException("Property 'dataSource' is required");      }      if (!isLazyInit()) {          getExceptionTranslator();      }  }

通过上面代码来看说白了就是用来初始化我们JdbcTemplate对象的,当然也做了不少操作,比如要判断是够有提供dataSource对象

第二步是调用对应的query方法,把sql和参数传入最后接收响应结果


创建完我们的jdbcTemplate对象后那么接下来才真正的进入到我们执行查询的query方法中~
JdbcTemplate.java

[Java] 纯文本查看 复制代码

?

01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

public <T> List<T> query(String sql, RowMapper<T> rowMapper, Object... args) throws DataAccessException {     //---------     //标注五     return (List)this.query((String)sql, (Object[])args, (ResultSetExtractor)(new RowMapperResultSetExtractor(rowMapper))); }   ...   public <T> T query(String sql, Object[] args, ResultSetExtractor<T> rse) throws DataAccessException {     return this.query(sql, this.newArgPreparedStatementSetter(args), rse); }   ...   public <T> T query(String sql, PreparedStatementSetter pss, ResultSetExtractor<T> rse) throws DataAccessException {     return this.query((PreparedStatementCreator)(new JdbcTemplate.SimplePreparedStatementCreator(sql)), (PreparedStatementSetter)pss, (ResultSetExtractor)rse); }   ...   //--------- //标注一 public <T> T query(PreparedStatementCreator psc, final PreparedStatementSetter pss, final ResultSetExtractor<T> rse) throws DataAccessException {     Assert.notNull(rse, "ResultSetExtractor must not be null");     this.logger.debug("Executing prepared SQL query");     //---------     //标注二     return this.execute(psc, new PreparedStatementCallback<T>() {         public T doInPreparedStatement(PreparedStatement ps) throws SQLException {             //---------             //标注三               ResultSet rs = null;               Object var4;             try {                 if (pss != null) {                     //将预编译的PreparedStatement对象和sql语句的?占位符和实际参数进行绑定                     pss.setValues(ps);                 }                 //获取结果集                 rs = ps.executeQuery();                 ResultSet rsToUse = rs;                 if (JdbcTemplate.this.nativeJdbcExtractor != null) {                     rsToUse = JdbcTemplate.this.nativeJdbcExtractor.getNativeResultSet(rs);                 }                 //---------                 //标注四                 var4 = rse.extractData(rsToUse);             } finally {                 JdbcUtils.closeResultSet(rs);                 if (pss instanceof ParameterDisposer) {                     ((ParameterDisposer)pss).cleanupParameters();                 }               }               return var4;         }     }); }

通过上面代码来看,我们很容易就能够追到标注一的地方,而这里也是触发查询操作的开始,接下来咱们则来对这个方法进行一个详细的剖析,看到标注二这里相信大家就能感到很眼熟了,因为上次我们在追update方法的源码时最终也会执行到execute方法,在这我们就直接将execute方法执行原理不在复述一遍了.我们直接从标注三开始,因为doInPreparedStatement方法的返回值其实就是最终的返回值

标注四
首先先来看下rse对象
    ResultSetExtractor<T> rse
这个对象是在标注五的地方进行的初始化,(ResultSetExtractor)(new RowMapperResultSetExtractor(rowMapper)),通过这段代码我们可以看到我们将RowMapperResultSetExtractor对象强制转换成ResultSetExtractor,但是创建RowMapperResultSetExtractor对象的时候我们传入了一个rowMapper对象,这个对象,则是咱们查询的时候传入的BeanPropertyRowMapper,接下来咱们从RowMapperResultSetExtractor开始分析

标注五
RowMapperResultSetExtractor

[Java] 纯文本查看 复制代码

?

01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

public class RowMapperResultSetExtractor<T> implements ResultSetExtractor<List<T>> {     private final RowMapper<T> rowMapper;     private final int rowsExpected;       public RowMapperResultSetExtractor(RowMapper<T> rowMapper) {         this(rowMapper, 0);     }       public RowMapperResultSetExtractor(RowMapper<T> rowMapper, int rowsExpected) {         Assert.notNull(rowMapper, "RowMapper is required");         this.rowMapper = rowMapper;         this.rowsExpected = rowsExpected;     }     //---------     //标注六     public List<T> extractData(ResultSet rs) throws SQLException {         List<T> results = this.rowsExpected > 0 ? new ArrayList(this.rowsExpected) : new ArrayList();         int var3 = 0;           while(rs.next()) {             results.add(this.rowMapper.mapRow(rs, var3++));         }           return results;     } }

通过上面的代码我们看到该类中保存了RowMapper和rowsExpected,因为走的是一个参数的构造所以rowsExpected为0,因此标注4中的rse对象的实际对象为RowMapperResultSetExtractor,而rse.extractData(rsToUse);方法则进入

标注六
    List<T> results = this.rowsExpected > 0 ? new ArrayList(this.rowsExpected) : new ArrayList();
经过之前的分析,rowsExpected默认为0,所以执行三元表达式的后半部分,创建一个新的ArrayList
    while(rs.next()) {
        //---------        
        //标注7
        results.add(this.rowMapper.mapRow(rs, var3++));
    }
这段代码中则是将结果集的每条结果添加到目标ArrayList中,因此我们又把注意力转向了this.rowMapper.mapRow(rs, var3++)

标注七
rowMapper这个对象实际类型为BeanPropertyRowMapper
BeanPropertyRowMapper初始化走的构造

[Java] 纯文本查看 复制代码

?

01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

public BeanPropertyRowMapper(Class<T> mappedClass) {     this.initialize(mappedClass); } protected void initialize(Class<T> mappedClass) {     this.mappedClass = mappedClass;     this.mappedFields = new HashMap();     this.mappedProperties = new HashSet();     PropertyDescriptor[] pds = BeanUtils.getPropertyDescriptors(mappedClass);     PropertyDescriptor[] var3 = pds;     int var4 = pds.length;       for(int var5 = 0; var5 < var4; ++var5) {         PropertyDescriptor pd = var3[var5];         if (pd.getWriteMethod() != null) {             this.mappedFields.put(this.lowerCaseName(pd.getName()), pd);             String underscoredName = this.underscoreName(pd.getName());             if (!this.lowerCaseName(pd.getName()).equals(underscoredName)) {                 this.mappedFields.put(underscoredName, pd);             }               this.mappedProperties.add(pd.getName());         }     }   }

上面用到了内省,这块简单介绍下,就是通过内省我们可以轻松获取到一个类的属性和相关get/set方法,只要属性和相关方法对象,那么我们就可以通过反射的形式将结果集中的数据封装到该对象中,而上面代码只是进行了参数的初始化方便mapRow方法进行封装而已

[Java] 纯文本查看 复制代码

?

01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59

public T mapRow(ResultSet rs, int rowNumber) throws SQLException {     Assert.state(this.mappedClass != null, "Mapped class was not specified");     //---------     //标注八:初始化目标对象     T mappedObject = BeanUtils.instantiate(this.mappedClass);     BeanWrapper bw = PropertyAccessorFactory.forBeanPropertyAccess(mappedObject);     this.initBeanWrapper(bw);     //结果集元数据     ResultSetMetaData rsmd = rs.getMetaData();     //获取一共多少列     int columnCount = rsmd.getColumnCount();     Set<String> populatedProperties = this.isCheckFullyPopulated() ? new HashSet() : null;     //---------     //标注九:初始化目标对象     //循环遍历每一列,将每列的值封装到实体bean的指定属性中     for(int index = 1; index <= columnCount; ++index) {         String column = JdbcUtils.lookupColumnName(rsmd, index);         String field = this.lowerCaseName(column.replaceAll(" ", ""));         PropertyDescriptor pd = (PropertyDescriptor)this.mappedFields.get(field);         if (pd == null) {             if (rowNumber == 0 && this.logger.isDebugEnabled()) {                 this.logger.debug("No property found for column '" + column + "' mapped to field '" + field + "'");             }         } else {             try {                 //获取该列的数据,同时找到判断具体数据类型                 Object value = this.getColumnValue(rs, index, pd);                 if (rowNumber == 0 && this.logger.isDebugEnabled()) {                     this.logger.debug("Mapping column '" + column + "' to property '" + pd.getName() + "' of type [" + ClassUtils.getQualifiedName(pd.getPropertyType()) + "]");                 }                   try {                     //将Bean的属性和获取到的值进行绑定                     bw.setPropertyValue(pd.getName(), value);                 } catch (TypeMismatchException var14) {                     if (value != null || !this.primitivesDefaultedForNullValue) {                         throw var14;                     }                       if (this.logger.isDebugEnabled()) {                         this.logger.debug("Intercepted TypeMismatchException for row " + rowNumber + " and column '" + column + "' with null value when setting property '" + pd.getName() + "' of type [" + ClassUtils.getQualifiedName(pd.getPropertyType()) + "] on object: " + mappedObject, var14);                     }                 }                   if (populatedProperties != null) {                     populatedProperties.add(pd.getName());                 }             } catch (NotWritablePropertyException var15) {                 throw new DataRetrievalFailureException("Unable to map column '" + column + "' to property '" + pd.getName() + "'", var15);             }         }     }       if (populatedProperties != null && !populatedProperties.equals(this.mappedProperties)) {         throw new InvalidDataAccessApiUsageException("Given ResultSet does not contain all fields necessary to populate object of class [" + this.mappedClass.getName() + "]: " + this.mappedProperties);     } else {         return mappedObject;     } }

标注八

[Java] 纯文本查看 复制代码

?

01 02 03 04 05 06 07 08 09 10 11 12 13 14

public static <T> T instantiate(Class<T> clazz) throws BeanInstantiationException {     Assert.notNull(clazz, "Class must not be null");     if (clazz.isInterface()) {         throw new BeanInstantiationException(clazz, "Specified class is an interface");     } else {         try {             return clazz.newInstance();         } catch (InstantiationException var2) {             throw new BeanInstantiationException(clazz, "Is it an abstract class?", var2);         } catch (IllegalAccessException var3) {             throw new BeanInstantiationException(clazz, "Is the constructor accessible?", var3);         }     } }

ok,到此为止我们就能分析出来,在这段for循环中才是将实体Bean对象中的属性值和rs中每列的数据进行绑定,从而完成将查询到的数据封装到Bean的过程.当然,这块因为篇幅的原因可能有点粗糙,但是主要起到一个抛砖引玉的作用~~~