简体   繁体   中英

How to map collections in Dozer

I'd like to do something like:

ArrayList<CustomObject> objects = new ArrayList<CustomObject>();
...
DozerBeanMapper MAPPER = new DozerBeanMapper();
...
ArrayList<NewObject> newObjects = MAPPER.map(objects, ...); 

Assuming:

<mapping>
  <class-a>com.me.CustomObject</class-a>
  <class-b>com.me.NewObject</class-b>   
    <field>  
      <a>id</a>  
      <b>id2</b>  
    </field>  
</mapping>

I tried :

ArrayList<NewObject> holder = new ArrayList<NewObject>();
MAPPER.map(objects, holder);

but the holder object is empty. I also played with changing the second argument without any luck...

To quote:

"Nested collections are handled automatically, but you are correct that top level collections need to be iterated over. Currently there isn't a more elegant way to handle this."

Someone has figured a way to do it without a looping construct in your code base , but I think it's just easier (and more readable/maintainable) to put it in your code. Hopefully they'll add this ability sooner than later.

I faced a similar issue, and decided on using a generic utility method to avoid iterating every time I needed to perform such mapping.

public static <T, U> List<U> map(final Mapper mapper, final List<T> source, final Class<U> destType) {

    final List<U> dest = new ArrayList<>();

    for (T element : source) {
        dest.add(mapper.map(element, destType));
    }

    return dest;
}

Usage would then be something like:

    final List<CustomObject> accounts..... 
    final List<NewObject> actual = Util.map(mapper, accounts, NewObject.class);

Possibly this could be simplified further though.

What is happening is that you are getting bitten by type erasure. At runtime, java only sees an ArrayList.class . The type of CustomObject and NewObject aren't there, so Dozer is attempting to map a java.util.ArrayList , not your CustomObject to NewObject .

What should work (totally untested):

List<CustomObject> ori = new ArrayList<CustomObject>();
List<NewObject> n = new ArrayList<NewObject>();
for (CustomObject co : ori) {
    n.add(MAPPER.map(co, CustomObject.class));
}

you can do it like this :

public <T,S> List<T> mapListObjectToListNewObject(List<S> objects, Class<T> newObjectClass) {
final List<T> newObjects = new ArrayList<T>();
for (S s : objects) {
    newObjects.add(mapper.map(s, newObjectClass));
}
return newObjects;

}

and use it :

ArrayList<CustomObject> objects = ....
List<NewObject> newObjects = mapListObjectToListNewObject(objects,NewObject.class);

I have done it using Java 8 and dozer 5.5. You don't need any XML files for mapping. You can do it in Java.

You don't need any additional mapping for lists , only thing you need is

you need to add the list as a field in the mapping

. See the sample bean config below.

Spring configuration class

@Configuration
public class Config {

@Bean
    public DozerBeanMapper dozerBeanMapper() throws Exception {
        DozerBeanMapper mapper = new DozerBeanMapper();
        mapper.addMapping( new BeanMappingBuilder() {
            @Override
            protected void configure() {
                mapping(Answer.class, AnswerDTO.class);
                mapping(QuestionAndAnswer.class, QuestionAndAnswerDTO.class).fields("answers", "answers");                  
            }
        });
        return mapper;
    }

}

//Answer class and AnswerDTO classes have same attributes

public class AnswerDTO {

    public AnswerDTO() {
        super();
    }

    protected int id;
    protected String value;

    //setters and getters
}

//QuestionAndAnswerDTO class has a list of Answers

public class QuestionAndAnswerDTO {

    protected String question;
    protected List<AnswerDTO> answers;

   //setters and getters
}

//LET the QuestionAndAnswer class has similar fields as QuestionAndAnswerDTO

//Then to use the mapper in your code, autowire it

@Autowired
private DozerBeanMapper dozerBeanMapper;
// in your method


 QuestionAndAnswerDTO questionAndAnswerDTO =
    dozerBeanMapper.map(questionAndAnswer, QuestionAndAnswerDTO.class);

Hope this will help someone follow the Java approach instead of XML.

For that use case I once wrote a little helper class:

import java.util.Collection;

/**
 * Helper class for wrapping top level collections in dozer mappings.
 * 
 * @author Michael Ebert
 * @param <E>
 */
public final class TopLevelCollectionWrapper<E> {

    private final Collection<E> collection;

    /**
     * Private constructor. Create new instances via {@link #of(Collection)}.
     * 
     * @see {@link #of(Collection)}
     * @param collection
     */
    private TopLevelCollectionWrapper(final Collection<E> collection) {
        this.collection = collection;
    }

    /**
     * @return the wrapped collection
     */
    public Collection<E> getCollection() {
        return collection;
    }

    /**
     * Create new instance of {@link TopLevelCollectionWrapper}.
     * 
     * @param <E>
     *            Generic type of {@link Collection} element.
     * @param collection
     *            {@link Collection}
     * @return {@link TopLevelCollectionWrapper}
     */
    public static <E> TopLevelCollectionWrapper<E> of(final Collection<E> collection) {
        return new TopLevelCollectionWrapper<E>(collection);
    }
}

You then would call dozer in the following manner:

private Mapper mapper;

@SuppressWarnings("unchecked")
public Collection<MappedType> getMappedCollection(final Collection<SourceType> collection) {
    TopLevelCollectionWrapper<MappedType> wrapper = mapper.map(
            TopLevelCollectionWrapper.of(collection),
            TopLevelCollectionWrapper.class);

    return wrapper.getCollection();
}

Only drawback: You get a "unchecked" warning on mapper.map(...) because of Dozers Mapper interface not handling generic types.

Not really an improvement, more like a syntactic sugar that can be achieved thanks to Guava (and most likely similar thing is possible with Apache Commons ):

final List<MyPojo> mapped = Lists.newArrayList(Iterables.transform(inputList, new Function<MyEntity, MyPojo>() {
    @Override public MyPojo apply(final MyEntity arg) {
        return mapper.map(arg, MyPojo.class);
    }
}));

This can also be turned into a generic function - as suggested in other answers.

You can implement your own mapper class which will extend dozer mapper. Example: Create a interface that adds additional method to dozer mapper:

public interface Mapper extends org.dozer.Mapper {
    <T> List<T> mapAsList(Iterable<?> sources, Class<T> destinationClass);
}

Next step: Write your own Mapper class by implementing above interface.

add below method to your implementation class:

public class MyMapper implements Mapper {
    @Override
    public <T> List<T> mapAsList(Iterable<?> sources, Class<T> destinationClass) {
        //can add validation methods to check if the object is iterable
        ArrayList<T> targets = new ArrayList<T>();
        for (Object source : sources) {
            targets.add(map(source, destinationClass));
        }
        return targets;
    }
    //other overridden methods.
}

Hope this helps

The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.

 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM