Wrote test, use mockito...drink mockito!

Hi all!

In our project we have parts that are written by our outsource partners. We are using their services in our code. Services injected from spring to our beans.

Yesterday, they made a weekly commit, and I've found out that my tests were broken :( One of their developer had commented my stub that implemented their interface. They said that he didn't want to change any code although I needed my stub to collect and analyze input in it.

I started solving this problem learning about mock objects. On my previous work, one of my coworkers used mockito and he liked it a lot :) I've found the information about it here: site. I suppose it's good that the last version was released in may. Also, I like projects that have good documentation and have *-all.jar. Mockito has all of that!

I needed only one method from two interfaces. And I've written simple mock objects and inject this mocks in spring.

    SomeService mockService = Mockito.mock(SomeService.class);
    Mockito.doAnswer(new Answer() {
        @Override
        public Object answer(InvocationOnMock invocation) throws Throwable {
            String[] users = (String[])invocation.getArguments()[0];
            String title = (String)invocation.getArguments()[1];
            //Collect information here.
            return null;
        }
    }).when(mockService).
            someMethod(Mockito.<string[]>any(), Mockito.anyString());

Also, I have got some troubles when injecting this service in spring: I could not do something like this:

myService.setSomeService(mockService);

because I am using @Transactional and having Proxy object. After some google searches I've found solution for this(I think, that it's not ideal, because I found some info about @Configurable beans and wrote TODO ;))

((MyServiceImpl)( ((Advised)myService).getTargetSource().getTarget())).setSomeService(mockService);

I commited this and...drink mockito :)!

Increase perfomance for insert operations to database

Hi all!

In this post, I would like to tell about the way to optimize INSERT operation to database.


In our project we use postgresql 8.3. I have big problems with speed on some operations. When I profiled the application, I find out that the biggest amount of time spent for JDBC operations. We getting data from external systems, and often we have more than 300000 inserts to table. For work with database we use JDBCTemplate from Spring framework.

All that you should do: realize class that implements BatchPreparedStatementSetter. I create universal abstract class that implements this interface.

abstract public class AbstractPreparedStatementSetter implements PreparedStatementSetter, BatchPreparedStatementSetter {
 private List objects;

 protected AbstractPreparedStatementSetter(List objects) {
  this.objects = objects;
 }

 public void setValues(PreparedStatement preparedStatement) throws SQLException {
  fillStatement(objects.get(0), preparedStatement);
 }

 @Override
 public void setValues(PreparedStatement preparedStatement, int i) throws SQLException {
  fillStatement(objects.get(i), preparedStatement);
 }

 @Override
 public int getBatchSize() {
  return objects.size();
 }

 abstract public void fillStatement(T object, PreparedStatement statement) throws SQLException;
}

Concrete implementation example:

public class RegistrationEntrySetter extends AbstractPreparedStatementSetter {
 public RegistrationEntrySetter(List objects) {
  super(objects);
 }

 @Override
 public void fillStatement(RegistrationEntry entry, PreparedStatement statement) throws SQLException {
  long time = entry.getActionDate() == null ? System.currentTimeMillis() : entry.getActionDate().getTime();
  statement.setString(1, entry.getLogin());
  statement.setTimestamp(2, new Timestamp(time));
  statement.setShort(3, entry.getActionResult());
  statement.setString(4, entry.getDeviceName());
  String comment = entry.getComment();
  if(comment.length() > 8190)
   comment = comment.substring(0, 8190);
  statement.setString(5, comment);
  statement.setShort(6, entry.getAcessLevel().getLevel());
  statement.setString(7, entry.getObjectType());
  statement.setString(8, entry.getObjectCategory());
  statement.setString(9, entry.getObjectId());
  statement.setInt(10, entry.getEventType().getId());
  statement.setString(11, entry.getActionName());
  statement.setString(12, entry.getUserName());
  statement.setShort(13, entry.isNotification() ? (short)1 : (short)0);
  statement.setString(14, entry.getRemoteAddress());
  statement.setString(15, PasswordEncoder.encode(entry.toString()));
 }
}

In spring xml context, I describe all implementations for statements setters for my classes.

    <bean id="preparedStatementsSetter" class="java.util.HashMap">
        <constructor-arg>
            <map>
                <entry key="com.blogspot.RegistrationEntry">
                    <value>com.otr.security.server.audit.sql.RegistrationEntrySetter</value>
                </entry>
                .....
                .....
            </map>
        </constructor-arg>
    </bean>

And in my dao I can get setter for concrete class for list of objects.
In this way we have one advantage, all setters are singleton and we have reusable prepared statements.

The sample dao may look like this:

public class AbstractDAO {
 //spring fields
 private SimpleJdbcTemplate template;
 private Map<Class, Class<? extends AbstractPreparedStatementSetter>> statementSetters;
 //spring fields

 public int[] batchUpdate(String sql, List objects) {
  if (!objects.isEmpty()) {
   try {
    AbstractPreparedStatementSetter setter = getSetter(objects, objects.get(0).getClass());
    if (setter != null) {
     return template.getJdbcOperations().batchUpdate(sql, setter);
    }
   } catch (Exception e) {
    log.error("Error while construct sql setter", e);
   }
  }

  return new int[]{};
 }

 private AbstractPreparedStatementSetter getSetter(List objects, Class objectClazz)
   throws InvocationTargetException, IllegalAccessException, InstantiationException, NoSuchMethodException {
  Class<? extends AbstractPreparedStatementSetter> setterClazz = statementSetters.get(objectClazz);

  if (setterClazz != null) {
   Constructor<? extends AbstractPreparedStatementSetter> c = setterClazz.getConstructor(List.class);
   return c.newInstance(objects);
  } else {
   log.warn("======================NO SQL SETTER FOR CLASS " + objectClazz.getName() + "======================");
   return null;
  }
 }
}

Placeholder trick in Spring

Hello everyone! In the Spring Framework (I use Spring 2.5), we can use property placeholders in the spring xml-based configuration file. I wrote simple example that consist of four beans. Two of them are simple beans, one is a factory and one is a result bean that we want to get from factory bean.

<?xml version="1.0" encoding="UTF-8"?>

<beans xmlns="http://www.springframework.org/schema/beans"

xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"

xsi:schemaLocation="http://www.springframework.org/schema/beans http://www.springframework.org/schema/beans/spring-beans-2.0.xsd">

    <bean id=”simpleFactory” class=”c om.blogger.dooman87.SimpleFactory”>

        <constructor-arg value=”$ {com.blogger.dooman87.resultBeanName}”>

    </bean>

    <bean id="resultBean" factory-bean=”s impleFactory”/>

    <bean id="simpleBeanFirst"

        class="com.blogger.dooman87.SimpleBeanFirst"/>

    <bean id="simpleBeanSecond"

        class="com.blogger.dooman87.SimpleBeanSecond"/>

</beans>

In this case, we should write additional class com.blogger.dooman87.SimpleFactory, that would instantiate result by first constructor argument. This class we can extend from AbstractFactoryBean. But as well we are able to replace the factory! This simple trick you can see below:

<?xml version="1.0" encoding="UTF-8"?>

<beans xmlns="http://www.springframework.org/schema/beans"

xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"

xsi:schemaLocation="http://www.springframework.org/schema/beans http://www.springframework.org/schema/beans/spring-beans-2.0.xsd">

    <alias name="${com.blogger.dooman87.resultBeanName}" alias="resultBean"/>

    <bean id="simpleBeanFirst"

        class="com.blogger.dooman87.SimpleBeanFirst"/>

    <bean id="simpleBeanSecond"

        class="com.blogger.dooman87.SimpleBeanSecond"/>

</beans>

We created alias for result bean, where the name is a placeholder value. That's all. Hope this post was interesting or you.