基于Spring的单元测试框架搭建及分布式数据源单元测试

阅读更多

 

在Space 1st Apps的项目中,我们使用了基于Spring的单元测试 ,  并结合DBUnit对数据库应用程序做单元测试。

其中,数据源我们使用了Oracle、MySql数据库,Oracle数据源只有一个数据库,而MySQL数据源是一个由多台MySQL数据库组成的分布式多数据源。

 

对于单数据源Single DataSource的单元测试用例,如果你需要在测试用例运行之前准备一些种子数据,那么,你只需要简单地继承抽象类:AbstractDataSourceTest,例如:

@ContextConfiguration(locations = { "/spring/space-apps-poke.xml" })
public class PokeInfoManagerTest extends AbstractDataSourceTest {
    ……
	@Override
	protected String getSeedXmlFile() {
		return "/seeds/poke/pokeinfo.xml";
	}
    ……
}

 

其中,@ContextConfiguration的参数locations的值,表示类路径下的Spring配置文件的位置,例如:/spring/space-apps-poke.xml,在开发时,你将该文件放在了src/conf文件下面,将该文件夹作为源文件夹使用,这样该文件夹下面的所有文件,就都会在类路径下面了,如下图所示:



必须实现的抽象方法:

protected abstract String getSeedXmlFile()

该方法实现,需要你返回类路径下面的测试种子数据XML文件的位置,测试种子数据文件统一放到src/conf.test下面,如下图所示:



 在测试用例中,如果需要引用Spring的bean,只需像下面代码这样:

……
@Autowired
private PokeInfoManager pokeInfoManager;
……
 

 一旦加上了@Autowired注解之后,id为“pokeInfoManager“的bean,就会自动注入到测试用例中。

 测试方法,只需在普通的方法前面加上注解@Test,就可以被单元测试框架运行,例如:

 

	@Test
	public void getAllPokeInfoNotCached() {
		memCachedTemplate.delete(AppsKeyAccessor.Memcached.SPACE_APPS_POKE);
		List pokeInfoList = pokeInfoManager.getAllPokeInfo();
		assertNotNull(pokeInfoList);
		assertTrue(!pokeInfoList.isEmpty());
	}
   

 

附上抽象超类的源码

@RunWith(SpringJUnit4ClassRunner.class)
@TestExecutionListeners( { DependencyInjectionTestExecutionListener.class,
		DirtiesContextTestExecutionListener.class })
@ContextConfiguration(locations = { "/spring/space-memcached.xml",
		"/spring/space-datasource.xml", "/spring/space-commons.xml",
		"/spring/space-datasource-multi.xml" })
public abstract class AbstractDataSourceTest {

	private Resource resource;
	private IDatabaseTester tester;
	@Autowired
	@Qualifier("spaceOracleDS")
	private DataSource dataSource;

	/**
	 * Returns the test data seeds file in XML format
	 */
	protected abstract String getSeedXmlFile();

	/**
	 * Creates a new IDatabaseTester.
* Default implementation returns a {@link DataSourceDatabaseTester} * configured with the value returned from {@link getDataSource()}. */ protected IDatabaseTester newDatabaseTester() { return new DataSourceDatabaseTester(getDataSource()); } /** *Returns the test DataSource. */ protected DataSource getDataSource() { return this.dataSource; } /** * Gets the IDatabaseTester for this testCase.
* If the IDatabaseTester is not set yet, this method calls * newDatabaseTester() to obtain a new instance. * * @throws Exception */ protected IDatabaseTester getDatabaseTester() throws Exception { if (this.tester == null) { this.tester = newDatabaseTester(); } return this.tester; } /** * Returns the database operation executed in test setup. default is * DatabaseOperation.REFRESH */ protected DatabaseOperation getSetUpOperation() throws Exception { return DatabaseOperation.REFRESH; } /** * Returns the database operation executed in test cleanup.default is * DatabaseOperation.NONE */ protected DatabaseOperation getTearDownOperation() throws Exception { return DatabaseOperation.NONE; } /** * Returns data set resource in spring resource mode. eg: /seeds/xxx.xml If * the Resource is not set yet, this method calls newDataSetResource() to * obtain a new instance. */ private Resource getDataSetResource() { if (this.resource == null) { this.resource = newDataSetResource(); } return this.resource; } /** * Creates a new Resource.
* Default implementation returns a {@link ClassPathResource} configured * with the value returned from {@link getSeedXmlFile()}. */ protected Resource newDataSetResource() { return new ClassPathResource(getSeedXmlFile()); } /** * Returns the test dataset. default is FlatXmlDataSet */ private IDataSet getDataSet() throws Exception { return new FlatXmlDataSet(getDataSetResource().getInputStream()); } @Before public void setUp() throws Exception { final IDatabaseTester databaseTester = getDatabaseTester(); databaseTester.setSetUpOperation(getSetUpOperation()); databaseTester.setDataSet(getDataSet()); databaseTester.onSetup(); } @After public void tearDown() throws Exception { final IDatabaseTester databaseTester = getDatabaseTester(); databaseTester.setTearDownOperation(getTearDownOperation()); databaseTester.setDataSet(getDataSet()); databaseTester.onTearDown(); tester = null; } }
 

 

如果你不需要测试种子数据,请不要继承AbstractDataSourceTest类,而是仅需继承Spring 测试框架自带的类:AbstractJUnit4SpringContextTests

 

如果你的测试用例想使用事务,譬如说,你不想每次运行测试用例,都把数据真实地插入到数据库中,你可以使用事务的自动回滚机制,这很简单,你只需要继承AbstractTransactionalDataSourceTest

@ContextConfiguration(locations = { "/spring/space-apps-album.xml" })
public class BoPhotoManagerTest extends AbstractTransactionalDataSourceTest {

    ……
	@Override
	protected String getSeedXmlFile() {
		return "/seeds/admin/bophoto/bophoto.xml";
	}
    ……
}
 

 以上代码和无事务支持,需要测试种子数据测试用例比较类似,在此,不再累赘。

 

 附上AbstractTransactionalDataSourceTest的源代码:

@TestExecutionListeners( { TransactionalTestExecutionListener.class })
@Transactional
@TransactionConfiguration(transactionManager = "jtaTransactionManager")
public abstract class AbstractTransactionalDataSourceTest extends
		AbstractDataSourceTest {
}
 

其中,注解@TransactionConfiguration的参数transactionManager的值为jtaTransactionManager,是Spring的bean的Id

 

如果测试用例需要多数据源分布式数据库,大致类似,简单说一下

 

对于需要测试种子数据的,请继承AbstractDistributedDataSourceTest,附上该类源码:

@RunWith(SpringJUnit4ClassRunner.class)
@TestExecutionListeners( { DependencyInjectionTestExecutionListener.class,
		DirtiesContextTestExecutionListener.class })
@ContextConfiguration(locations = { "/spring/space-memcached.xml",
		"/spring/space-datasource.xml", "/spring/space-commons.xml",
		"/spring/space-datasource-multi.xml" })
public abstract class AbstractDistributedDataSourceTest {

	private Resource resource;
	private IDistributedDatabaseTester tester;
	@Autowired
	@Qualifier("distributeDBDataSource")
	private DataSource dataSource;

	/**
	 * Returns the test data seeds file in XML format
	 */
	protected abstract String getSeedXmlFile();

	/**
	 * Creates a new IDistributedDatabaseTester.
* Default implementation returns a * {@link DistributedDataSourceDatabaseTester} configured with the value * returned from {@link getDataSource()}. */ protected IDistributedDatabaseTester newDatabaseTester() { return new DistributedDataSourceDatabaseTester(getDataSource()); } /** *Returns the test DataSource. */ protected DataSource getDataSource() { return this.dataSource; } /** * Gets the IDatabaseTester for this testCase.
* If the IDatabaseTester is not set yet, this method calls * newDatabaseTester() to obtain a new instance. * * @throws Exception */ protected IDistributedDatabaseTester getDatabaseTester() throws Exception { if (this.tester == null) { this.tester = newDatabaseTester(); } return this.tester; } /** * Returns the database operation executed in test setup. default is * DatabaseOperation.REFRESH */ protected DistributedDatabaseOperation getSetUpOperation() throws Exception { return DistributedDatabaseOperation.REFRESH; } /** * Returns the database operation executed in test cleanup.default is * DatabaseOperation.NONE */ protected DistributedDatabaseOperation getTearDownOperation() throws Exception { return DistributedDatabaseOperation.NONE; } /** * Returns data set resource in spring resource mode. eg: /seeds/xxx.xml If * the Resource is not set yet, this method calls newDataSetResource() to * obtain a new instance. */ private Resource getDataSetResource() { if (this.resource == null) { this.resource = newDataSetResource(); } return this.resource; } /** * Creates a new Resource.
* Default implementation returns a {@link ClassPathResource} configured * with the value returned from {@link getSeedXmlFile()}. */ protected Resource newDataSetResource() { return new ClassPathResource(getSeedXmlFile()); } /** * Returns the test dataset. default is FlatXmlDataSet */ private IExtDataSet getDataSet() throws Exception { return new ExtXmlDataSet(getDataSetResource().getInputStream()); } @Before public void setUp() throws Exception { final IDistributedDatabaseTester databaseTester = getDatabaseTester(); databaseTester.setSetUpOperation(getSetUpOperation()); databaseTester.setDataSet(getDataSet()); databaseTester.onSetup(); } @After public void tearDown() throws Exception { final IDistributedDatabaseTester databaseTester = getDatabaseTester(); databaseTester.setTearDownOperation(getTearDownOperation()); databaseTester.setDataSet(getDataSet()); databaseTester.onTearDown(); tester = null; } }
   

 如果需要事务回滚机制,请继承AbstractTransactionalDistributedDataSourceTest,附上该类源码

@TestExecutionListeners( { TransactionalTestExecutionListener.class })
@Transactional
@TransactionConfiguration(transactionManager = "jtaTransactionManager")
public abstract class AbstractTransactionalDistributedDataSourceTest extends
		AbstractDistributedDataSourceTest {

}
 

 

好了,到目前为止,测试用例是写好了,但是测试种子数据,还没有准备,这个XML文件,看起来还蛮复杂的,大家看一下样本,



	
	
	
	
	
	
	
	……

 

以上XML片段是一个样本,根节点是dataset,每个子节点表示表的名称,如FA_POKE_INFO,每个节点的属性如POKE_INFO_ID,表示数据库表的列名,每个节点的属性的值就是数据表对应列的值

 

OK,从头开始写这么一个文件,我觉得你一定会没有耐心,还是有一些工作量的。

 

还好,我写了几个工具类,你只需要简单地修改几个参数, 然后运行一下,上述的XML代码就会生成了,好,我们来看代码说话,由于我们使用了Oracle和MySQL数据库,所以我使用了模板方法模式,做了少许抽象,DatabaseExporter源码:

public abstract class DatabaseExporter {

	private IDatabaseConnection databaseConnection;
	private Connection jdbcConnection;

	/**
	 * write DTD file
	 * 
	 * @param dtdFileName
	 * 
	 * @throws Exception
	 */
	protected void writeDataDtd(String dtdFileName) throws Exception {
		IDataSet dataSet = getDatabaseConnection().createDataSet();
		Writer out = new OutputStreamWriter(new FileOutputStream(dtdFileName));
		FlatDtdWriter datasetWriter = new FlatDtdWriter(out);
		datasetWriter.setContentModel(FlatDtdWriter.CHOICE);
		datasetWriter.write(dataSet);
	}

	private IDatabaseConnection getDatabaseConnection() throws Exception {
		if (this.databaseConnection == null) {
			this.databaseConnection = newDatabaseConnection(getJdbcConnection());
		}
		return this.databaseConnection;
	}

	private Connection getJdbcConnection() throws Exception {
		if (this.jdbcConnection == null) {
			this.jdbcConnection = newJdbcConnection();
		}
		return this.jdbcConnection;
	}

	/**
	 * full database export
	 * 
	 * @param fullXmlFileName
	 * @param isFlatXml
	 * 
	 * @throws Exception
	 */
	protected void exportFullDataSet(String fullXmlFileName, boolean isFlatXml)
			throws Exception {
		IDataSet fullDataSet = getDatabaseConnection().createDataSet();
		if (isFlatXml) {
			FlatXmlDataSet.write(fullDataSet, new FileOutputStream(
					fullXmlFileName));
		} else {
			XmlDataSet
					.write(fullDataSet, new FileOutputStream(fullXmlFileName));
		}
	}

	/**
	 * dependent tables database export: export table X and all tables that have
	 * a PK which is a FK on X, in the right order for insertion
	 * 
	 * @param rootTableName
	 * @param dependentXmlFileName
	 * @param isFlatXml
	 * @throws Exception
	 */
	protected void exportDependentDataSet(String rootTableName,
			String dependentXmlFileName, boolean isFlatXml) throws Exception {
		String[] depTableNames = TablesDependencyHelper.getAllDependentTables(
				getDatabaseConnection(), rootTableName);
		IDataSet depDataset = getDatabaseConnection().createDataSet(
				depTableNames);
		if (isFlatXml) {
			FlatXmlDataSet.write(depDataset, new FileOutputStream(
					dependentXmlFileName));
		} else {
			XmlDataSet.write(depDataset, new FileOutputStream(
					dependentXmlFileName));
		}
	}

	/**
	 * partial database export
	 * 
	 * @param tableName
	 * @param partialXmlFileName
	 * @param isFlatXml
	 * @throws Exception
	 */
	protected void exportPartialDataSet(String tableName,
			String partialXmlFileName, boolean isFlatXml) throws Exception {
		QueryDataSet partialDataSet = new QueryDataSet(getDatabaseConnection());
		partialDataSet.addTable(tableName);
		if (isFlatXml) {
			FlatXmlDataSet.write(partialDataSet, new FileOutputStream(
					partialXmlFileName));
		} else {
			XmlDataSet.write(partialDataSet, new FileOutputStream(
					partialXmlFileName));
		}
	}

	/**
	 * partial multiple database export
	 * 
	 * @param tableName2Query
	 * @param partialXmlFileName
	 * @param isFlatXml
	 * @throws Exception
	 */
	protected void exportMultiPartialDataSet(
			Map tableName2Query, String partialXmlFileName,
			boolean isFlatXml) throws Exception {
		QueryDataSet partialDataSet = new QueryDataSet(getDatabaseConnection());
		Validate.notEmpty(tableName2Query,
				"Please check tableName2Query, it must not be null or empty");
		Map treeMap = new TreeMap(
				tableName2Query);
		Set> entrys = treeMap.entrySet();
		for (Entry entry : entrys) {
			if (StringUtils.isBlank(entry.getValue())) {
				partialDataSet.addTable(entry.getKey());
			} else {
				partialDataSet.addTable(entry.getKey(), entry.getValue());
			}
		}

		if (isFlatXml) {
			FlatXmlDataSet.write(partialDataSet, new FileOutputStream(
					partialXmlFileName));
		} else {
			XmlDataSet.write(partialDataSet, new FileOutputStream(
					partialXmlFileName));
		}

	}

	/**
	 * partial database export
	 * 
	 * @param tableName
	 * @param query
	 * @param isFlatXml
	 * @throws Exception
	 */
	protected void exportPartialDataSet(String tableName, String query,
			String partialXmlFileName, boolean isFlatXml) throws Exception {
		QueryDataSet partialDataSet = new QueryDataSet(getDatabaseConnection());
		partialDataSet.addTable(tableName, query);
		if (isFlatXml) {
			FlatXmlDataSet.write(partialDataSet, new FileOutputStream(
					partialXmlFileName));
		} else {
			XmlDataSet.write(partialDataSet, new FileOutputStream(
					partialXmlFileName));
		}
	}

	protected abstract Connection newJdbcConnection() throws Exception;

	protected abstract IDatabaseConnection newDatabaseConnection(
			Connection jdbcConnection);

}
 

 MySQL 数据导出工具源代码:

public class MysqlDatabaseExporter extends DatabaseExporter {

	@Override
	protected IDatabaseConnection newDatabaseConnection(
			Connection jdbcConnection) {
		return new MySqlConnection(jdbcConnection, null);
	}

	@Override
	protected Connection newJdbcConnection() throws Exception {
		Class.forName("com.mysql.jdbc.Driver");
		return DriverManager.getConnection(
				"jdbc:mysql://10.2.225.151:3306/SNS", "sns", "alisoft");
	}

	/**
	 * @param args
	 * @throws Exception
	 */
	public static void main(String[] args) throws Exception {
		MysqlDatabaseExporter exporter = new MysqlDatabaseExporter();
		Map map = new HashMap();
		map.put("FA_GIFT_RECEIVER",
				"SELECT * FROM SNS.FA_GIFT_RECEIVER F limit 0,10;");
		map.put("FA_GIFT_SENDER",
				"SELECT * FROM SNS.FA_GIFT_SENDER F limit 0,10;");
		map.put("FA_GIFT_STAT", "SELECT * FROM SNS.FA_GIFT_STAT F limit 0,10;");
		exporter.exportMultiPartialDataSet(map, "statefullGift.xml", false);
		// exporter.writeDataDtd("myFile.dtd");
		// exporter.exportFullDataSet("full.xml", false);
		// exporter.exportDependentDataSet("FA_PHOTO", "dependents.xml", false);
	}
}
 

 Oracle 数据导出工具源代码:

public class OracleDatabaseExporter extends DatabaseExporter {

	@Override
	protected IDatabaseConnection newDatabaseConnection(
			Connection jdbcConnection) {
		OracleConnection connection = new OracleConnection(jdbcConnection, null);
		connection.getConfig().setFeature(
				DatabaseConfig.FEATURE_SKIP_ORACLE_RECYCLEBIN_TABLES, true);
		return connection;
	}

	@Override
	protected Connection newJdbcConnection() throws Exception {
		Class.forName("oracle.jdbc.driver.OracleDriver");
		return DriverManager.getConnection(
				"jdbc:oracle:thin:@10.2.224.34:1521:aepdb", "sns", "sns");
	}

	/**
	 * @param args
	 * @throws Exception
	 * @throws Exception
	 */
	public static void main(String[] args) throws Exception {
		OracleDatabaseExporter exporter = new OracleDatabaseExporter();
		exporter.exportPartialDataSet("FA_POKE_INFO", "partial.xml", false);
		exporter.writeDataDtd("myFile.dtd");
		// exporter.exportFullDataSet("full.xml", false);
		exporter.exportDependentDataSet("BO_PHOTO", "dependents.xml", false);
	}
}
   

代码很简单,如果你熟悉DBUnit,大家一看就会明白,我主要说说怎么个用法,对了,有一个小插曲,我们的多数据源分布式数据库,是根据用户的longId进行取模,然后选择不同的数据库的,这样为多数据源分布式数据库准备单元测试的种子数据,就会产生很大的障碍,我们必须能够让种子数据自动选择他们应该归属的数据库。

 

DBUnit的测试种子数据,大家使用最多的应该是FlatXMLFile,如上面的测试种子数据例子

 

在插入测试种子数据时,我们无法知道XML节点的哪个字段是作为选择多数据源的字段,我们使用isResourceId表示,平行化的结构很难表达这个ResourceId的意思。还好,DBUnit有一种非平行化的XML结构,如下XML代码所示:



	POKE_IDSENDER_LONG_IDRECEIVER_LONG_IDPOKE_CONTENTGMT_CREATEGMT_MODIFIEDIS_DELETED
   

以上XML片段中,dataset表示根节点,table表示一个表,name属性表示table的名称,column表示table的一个列,中间的文本表示列名。

像以上这样,我们就可以在某个数据列上面表示哪个是ResoutceId,如上面例子中:isResourceId="true"的列,RECEIVIER_LONG_I,当然,默认的DBUnit是不支持这种扩展的,为了这个扩展,我费了很多的力气,感觉DBUnit的扩展性不是很好,至于扩展的代码,有很多,这里就不一一列举了。

 

说说上面的Exporter的使用方法,为了导出数据库中某个表,使用:

protected void exportPartialDataSet(String tableName,
            String partialXmlFileName, boolean isFlatXml)
protected void exportPartialDataSet(String tableName, String query,
			String partialXmlFileName, boolean isFlatXml) throws Exception
  • tableName,表示你要导出数据库的表名
  • query,表示针对tableName的一个select查询语句,如:"select * from tableName limit 0, 1000"
  • partialXmlFileName,表示导出的XML种子数据的文件名称
  • isFlatXml,表示导出的XML文件是否是Flat结构,true,表示Flat XML,false表示树状XML

 

为了导出数据库中的多个表数据到一个XML文件中,使用:

protected void exportMultiPartialDataSet(
			Map tableName2Query, String partialXmlFileName,
			boolean isFlatXml) throws Exception
 
  • tableName2Query,表示tableName到query语句的map,key为tableName,value为query语句,可以为空
  • partialXmlFileName,表示导出的XML种子数据的文件名称
  • isFlatXml,表示导出的XML文件是否是Flat结构,true,表示Flat XML,false表示树状XML

另外,还有两个方法,其一:

protected void exportFullDataSet(String fullXmlFileName, boolean isFlatXml)
			throws Exception
 

该方法用于导出数据库中所有表到一个XML文件中,一般情况下,不需要使用该方法,尤其是数据量很大的时候,参数定义如下:

  • fullXmlFileName,表示导出的XML种子数据的文件名称
  • isFlatXml,表示导出的文件是否是Flat结构,true,表示Flat XML,false表示树状XML

其二:

protected void writeDataDtd(String dtdFileName) throws Exception

该方法用于导出数据库表结构到一个DTD文件中,注意:该方法,可能对Oracle 10i 以上开启回收站功能的数据库,运行会失败,解决方案,就是禁用这个功能,需要DBA支持,遇到该问题时,你可以Google一下。

 

好了,说了这么多了,希望对你有用

  • 基于Spring的单元测试框架搭建及分布式数据源单元测试_第1张图片
  • 大小: 30 KB
  • 基于Spring的单元测试框架搭建及分布式数据源单元测试_第2张图片
  • 大小: 12.7 KB
  • 查看图片附件

你可能感兴趣的:(单元测试,框架,Spring,数据结构,SNS)