我有一个Java SE 8 Spring 4.1.6-RELEASE应用程序,我正在实现org.springframework.jdbc.core.RowMapper<T>
接口,我对java.sql.ResultSet
传递的T mapRow(ResultSet rs, int rowNum)
接口有一些疑问。方法。
当我检查ResultSet
类时,我看到了一堆方法来获取列值:
╔══════════════╦═════════════════════╦════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════╗ ║ Return Type ║ Method ║ Return (javadoc, se 8) ║ ╠══════════════╬═════════════════════╬════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════╣ ║ String ║ getString ║ the column value; if the value is SQL NULL, the value returned is null ║ ║ boolean ║ getBoolean ║ the column value; if the value is SQL NULL, the value returned is false ║ ║ byte ║ getByte ║ the column value; if the value is SQL NULL, the value returned is 0 ║ ║ short ║ getShort ║ the column value; if the value is SQL NULL, the value returned is 0 ║ ║ int ║ getInt ║ the column value; if the value is SQL NULL, the value returned is 0 ║ ║ long ║ getLong ║ the column value; if the value is SQL NULL, the value returned is 0 ║ ║ float ║ getFloat ║ the column value; if the value is SQL NULL, the value returned is 0 ║ ║ double ║ getDouble ║ the column value; if the value is SQL NULL, the value returned is 0 ║ ║ BigDecimal ║ getBigDecimal ║ the column value; if the value is SQL NULL, the value returned is null ║ ║ byte[] ║ getBytes ║ the column value; if the value is SQL NULL, the value returned is null ║ ║ Date ║ getDate ║ the column value; if the value is SQL NULL, the value returned is null ║ ║ Time ║ getTime ║ the column value; if the value is SQL NULL, the value returned is null ║ ║ Timestamp ║ getTimestamp ║ the column value; if the value is SQL NULL, the value returned is null ║ ║ InputStream ║ getAsciiStream ║ a Java input stream that delivers the database column value as a stream of one-byte ASCII characters; if the value is SQL NULL, the value returned is null ║ ║ Reader ║ getCharacterStream ║ a java.io.Reader object that contains the column value; if the value is SQL NULL, the value returned is null in the Java programming language ║ ║ InputStream ║ getBinaryStream ║ a Java input stream that delivers the database column value as a stream of uninterpreted bytes; if the value is SQL NULL, the value returned is null ║ ║ <T> T ║ getObject ║ an instance of type holding the column value ║ ╚══════════════╩═════════════════════╩════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════╝
通常的期望/做法是:
rs.getObject("COLUMN_NAME", Boolean.class);
rs.getObject("COLUMN_NAME", Byte.class);
rs.getObject("COLUMN_NAME", Short.class);
rs.getObject("COLUMN_NAME", Integer.class);
rs.getObject("COLUMN_NAME", Long.class);
等,对于所有原始类型?其他所有内容都会为null
的实例返回SQL NULL
。
如果是这样的话,当有类型的Object方法存在时,为不同类型设置所有方法的重点是什么?
此外,每种方法的优点/缺点是什么?
使用getInt(String columnLabel)
:
Integer resultingActionId = rs.getInt("RESULTING_ACTION_ID");
if (rs.wasNull) {
resultingActionId = null
}
使用getObject(String columnLabel)
并转换为Integer
:
Integer resultingActionId = (Integer) rs.getObject("RESULTING_ACTION_ID");
使用getObject(String columnLabel, Class type)
:
Integer resultingActionId = rs.getObject("RESULTING_ACTION_ID", Integer.class);
例如,我注意到org.springframework.jdbc.core.JdbcTemplate
曾经有 ,queryForLong
等方法来获取单个来自单行查询的值,并将它们全部替换为类型queryForInt
queryForObject
方法。
谢谢!
答案 0 :(得分:1)
如果您查看java.sql.ResultSet,您可以看到您不需要如此明确。实际上,除非你有一个用于连接的typeMapper,它允许你使用getObject方法,否则它将不起作用(java.sql.ResultSet.getObject)。
我不知道它是否会对你有所帮助,但我设法找到了我自己的RowMapper,它非常适合我的需求。
private class ShabaUserMapper implements RowMapper<ShabaUser>
{
@Override
public ShabaUser mapRow( ResultSet rs, int rowNum ) throws SQLException
{
Collection<SimpleGrantedAuthority> roles = new ArrayList<SimpleGrantedAuthority>();
String auths = rs.getString( "role" );
roles.add( new SimpleGrantedAuthority( auths ) );
ShabaUser user = new ShabaUser( rs.getString( "username" ), rs.getString( "password" ),
rs.getBoolean( "enabled" ), rs.getString( "first_name" ),
rs.getString( "last_name" ), rs.getString( "email" ),
rs.getString( "date_joined" ), rs.getString( "last_online" ), true, true, true,
roles );
// Can be null!
Integer awesomeness = rs.getInt( "awesomeness" );
if ( rs.wasNull() )
{
awesomeness = null;
}
user.setAwesomeness( awesomeness );
return user;
}
}
private class ShabaUserListExtractor implements ResultSetExtractor<List<ShabaUser>>
{
private final ShabaUserMapper rowMapper;
private int rowsExpected;
public ShabaUserListExtractor()
{
this( new ShabaUserMapper(), 0 );
}
public ShabaUserListExtractor( ShabaUserMapper rowMapper, int rowsExpected )
{
Assert.notNull( rowMapper, "RowMapper is required" );
this.rowMapper = rowMapper;
this.rowsExpected = rowsExpected;
}
@Override
public List<ShabaUser> extractData( ResultSet rs ) throws SQLException
{
HashMap<String, ShabaUser> results = ( this.rowsExpected > 0
? new HashMap<String, ShabaUser>(
rowsExpected )
: new HashMap<String, ShabaUser>() );
int rowNum = 0;
while ( rs.next() )
{
ShabaUser user = rowMapper.mapRow( rs, rowNum++ );
if ( results.containsKey( user.getUsername() ) )
{
ShabaUser inUser = results.get( user.getUsername() );
ArrayList<GrantedAuthority> combinedAuthorities = new ArrayList<GrantedAuthority>();
combinedAuthorities.addAll( inUser.getAuthorities() );
combinedAuthorities.addAll( user.getAuthorities() );
results.put( user.getUsername(),
createUserDetails( user.getUsername(), user, combinedAuthorities ) );
} else
{
results.put( user.getUsername(), user );
}
}
return new ArrayList<ShabaUser>( results.values() );
}
}
我意识到这是很多代码,但希望你能看到这里完成了什么。实际的RowMapper实现实际上意味着容纳从行信息中提取对象的所有“脏工作”。
只要您的数据库设置正确并且您使其成为必需列上的NOT NULL,您将永远不会遇到拔出空行的问题。虽然我认为从ResultSet检查null响应并不会有什么坏处,但是如果列应该有值,你仍然只会抛出异常。
答案 1 :(得分:1)
resultSet.getXXX(String columnLabel)
,则使用 SQL NULL
,或者您只关心获取原始类型(无对象、装箱或数组类型)resultSet.getObject(String columnLabel, Class type)
,则使用 SQL NULL
但如果 type 是装箱类型,则不要取消装箱,例如Integer
因为原语不能 NullPointerException
null
进行转换如果列不允许 resultSet.getObject(String columnLabel)
,或者您的代码严格需要或期望原始类型,即没有对象类型(这包括装箱类型,例如 SQL NULL
、{{3 }} 例如 Integer
) 然后使用特定的 get 方法:
byte[]
据我所知,这种方法的原因是方便,因为它可以优雅地处理 // Integer is 0 if value is SQL NULL
int i = resultSet.getInt(column);
// This is fine but Integer will never be null. This may or may not be what you want
Integer j = resultSet.getInt(column);
,因为 primitive arrays which are actually objects 例如为 SQL NULL
返回 0。
如果列确实允许 int
,或者您的代码需要对象类型(包括装箱类型或数组),则使用 SQL NULL
和类型的类,但不要在使用此方法时取消装箱类型:
resultSet.getObject
此方法对于受支持但没有特定 get 方法的其他 SQL 类型也很有用,即超出 JDBC 规范 primitives cannot be null
中指定的最小值:
// Integer will be null on SQL NULL but that's OK for boxed types
Integer i = resultSet.getObject(column, Integer.class);
// Throws NullPointerException on SQL NULL since primitives can't be null
int unbox = resultSet.getObject(column, Integer.class);
// Integer will be 0 since getInt returns 0 on SQL NULL
Integer autobox = resultSet.getInt(column);
但请注意,直到 Java 7 才添加此方法,因此如果您使用的是 Java 6,则需要对其进行强制转换:
// No getUUID method but it's supported by Postgres so no need for custom mapping
UUID uuid = resultSet.getObject(column, UUID.class);
所以除非你真的不知道或不关心返回类型,否则没有理由使用它,除非你喜欢它的风格,或者你需要在 Java 6 上运行。
最后,我的强烈建议是在使用(或拆箱)装箱类型时要谨慎,在使用 JDBC 时避免意外的 UUID uuid = (UUID) resultSet.getObject(column);
。要么避免在您的架构中使用 NullPointerException
,例如如果可以,默认为 0,如果不能确定,则明确检查/转换。