Force the usage of a JPA AttributeConverter for enums - java

We're trying to figure out a robust way of persisting enums using JPA. The common approach of using #Enumerated is not desirable, because it's too easy to break the mappings when refactoring. Each enum should have a separate database value that can be different than the enum name/order, so that you can safely change the name or internal ordering (e.g. the ordinal values) of the enum without breaking anything. E.g. this blog post has an example on how to achieve this, but we feel the suggested solution adds too much clutter to the code. We'd like to achieve a similar result by using the new AttributeConverter mechanism introduced in JPA 2.1. We have an interface that each enum should implement that defines a method for getting the value that is used to store the enum in the database. Example:
public interface PersistableEnum {
String getDatabaseValue();
}
...
public enum SomeEnum implements PersistableEnum {
FOO("foo"), BAR("bar");
private String databaseValue;
private SomeEnum(String databaseValue) {
this.databaseValue = databaseValue;
}
public void getDatabaseValue() {
return databaseValue;
}
}
We also have a base converter that has the logic for converting enums to Strings and vice versa, and separate concrete converter classes for each enum type (AFAIK, a fully generic enum converter is not possible to implement, this is also noted in this SO answer). The concrete converters then simply call the base class that does the conversion, like this:
public abstract class EnumConverter<E extends PersistableEnum> {
protected String toDatabaseValue(E value) {
// Do the conversion...
}
protected E toEntityAttribute(Class<E> enumClass, String value) {
// Do the conversion...
}
}
...
#Converter(autoApply = true)
public class SomeEnumConverter extends EnumConverter<SomeEnum>
implements AttributeConverter<SomeEnum, String> {
public String convertToDatabaseColumn(SomeEnum attribute) {
return toDatabaseValue(attribute);
}
public SomeEnum convertToEntityAttribute(String dbData) {
return toEntityAttribute(SomeEnum.class, dbData);
}
}
However, while this approach works very nicely in a technical sense, there's still a pretty nasty pitfall: Whenever someone creates a new enum class whose values need to be stored to the database, that person also needs to remember to make the new enum implement the PersistableEnum interface and write a converter class for it. Without this, the enum will get persisted without a problem, but the conversion will default to using #Enumerated(EnumType.ORDINAL), which is exactly what we want to avoid. How could we prevent this? Is there a way to make JPA (in our case, Hibernate) NOT default to any mapping, but e.g. throw an exception if no #Enumerated is defined on a field and no converter can be found for the type? Or could we create a "catch all" converter that is called for all enums that don't have their own specific converter class and always throw an exception from there? Or do we just have to suck it up and try to remember the additional steps each time?

You want to ensure that all Enums are instances of PersistableEnum.
You need to set a Default Entity Listener (an entity listener whose callbacks apply to all entities in the persistence unit).
In the Default Entity Listener class implement the #PrePersist method and make sure all the Enums are instances of PersistableEnum.

Related

How to distinguish between instances of class?

I have class Validator, which manage all validation criteria from files and database. But this criteria are loaded by Loader like this:
Validator validator = Loader.load("clients"); //get all from clients.cfg file
What is the best approach to determine from another class, which criteria are currently loaded?
Importer importer;
Validator clientsValidator = Loader.load("clients");
Validator addressValidator = Loader.load("address"); ...
importer.validate(data, clientsValidator, addressValidator);
public class Importer{
public void validate(Data data, Validator... validator){
...
validateClient(data, one of validators);
validateAddress(data, another of validator);
...
}
}
I need to know in Importer class, which Validator is for Clients, which for Addresses... Any good approaches?
The best way would be for you to be add a field and accompanying method to Validator to return the identifier (e.g. "clients") with which it was created.
Alternatively, if by using a different identifier when calling Loader.load() you get back instances of different classes implementing the Validator interface, then you can use the Object.getClass() method to tell those classes apart. If those classes are within a pretty small set you might even get away with using instanceof directly.
We would need more information, such as what Loader does exactly, what Validator is and how much you are allowed to change their code before being able to provide a more concrete answer.
EDIT:
Quite honestly, perhaps you should reconsider a redesign of your data model. As it stands, you can apparently mix clients and addresses without any checks. You should restructure your code to be able to rely on the type safety features of Java.
One way would be to have a generic class/interface Validator<T>, where T would the class of the validated objects:
public interface Validator<T> {
public boolean validate(T object);
}
You could then have specific Data subclasses for your data, such as Address or Client, and set typed Validator objects to Importer through specific methods:
public class Importer {
public void addAddressValidator(Validator<Address> validator) {
...
}
public void addClientValidator(Validator<Client> validator) {
...
}
}
This is far safer than mixing all validator objects in a single variadic method call, and it is also the preferred approach of most common frameworks in the wild.
Why not have a getSource() in Validator which gets set when Loader loads the source.
Thinking more about the specific question below :
I need to know in Importer class, which Validator is for Clients,
which for Addresses... Any good approaches?
Actually a better way to do this is if Loader can return a ClientValidator (implementation of Validator) for client and AddressValidator for addresses.
That way you can avoid the if-else conditions and directly call validate on the Validator class
Pass the validators by position. You must also check if the specific validator is null or not before you use.
public void validate(Data data,
Validator clientsValidator,
Validator addressValidator) {
...
if (clientsValidator != null) {
validateClient(data, clientsValidator);
}
if (addressValidator != null) {
validateAddress(data, addressValidator);
}
...
}

Efficient Java Annotations

I'm a huge fan of Java's annotations, but find it a pain in the neck to have to include Google's Reflections or Scannotations every time I want to make my own.
I haven't been able to find any documentation about Java being able to automatically scan for annotations & use them appropriately, without the help of a container or alike.
Question: Have I missed something fundamental about Java, or were annotations always designed such that manual scanning & checking is required? Is there some built-in way of handling annotations?
To clarify further
I'd like to be able to approach annotations in Java a little more programatically. For instance, say you wanted to build a List of Cars. To do this, you annotate the list with a class that can populate the list for you. For instance:
#CarMaker
List<Car> cars = new List<Car>();
In this example, the CarMaker annotation is approached by Java, who strikes a deal and asks them how many cars they want to provide. It's up to the CarMaker annotation/class to then provide them with a list of which cars to include. This could be all classes with #CarType annotations, and a Car interface.
Another way of looking at it, is that if you know you want to build something like this: List<Car> cars, you could annotate it with #ListMaker<Car>. The ListMaker is something built into Java. It looks for all classes annotated with #CarType, and populates the list accordingly.
You can create your own annotations and apply them to your own classes.
If you specify that an annotation is detectable at runtime, you can process it easily with reflection.
For example, you could use something like this to print the name of each field in a class that has been marked with the Funky annotation:
for (Field someField : AnnotatedClass.getClass().getDeclaredFields()) {
if (someField.isAnnotationPresent(Funky.class)) {
System.out.println("This field is funky: " + someField.getName());
}
}
The code to declare the Funky annotation would look something like this:
package org.foo.annotations;
#Retention(RetentionPolicy.RUNTIME)
#Target(ElementType.FIELD)
public #interface Funky { }
Here's a class that uses the annotation:
package org.foo.examples;
import org.foo.annotations.Funky;
public class AnnotatedClass {
#Funky
private String funkyString;
private String nonFunkyString;
#Funky
private Integer funkyInteger;
private Integer nonFunkyInteger;
}
Here's some more reading on Annotations.
Here are the javadocs for the classes used above:
Retention annotation
RetentionPolicy enum
Target annotation
Field class
isAnnotationPresent() method
getDeclaredFields() method
I'm trying to understand your car example, but I'm not sure I follow what you want.
If you had a list of objects (Jaguar, Porche, Ferrari, Kia) that extend Car and are marked with various car-related annotations, you could create an object that filters the list based on annotations.
The code might look like this:
#WorldsFinestMotorCar
class Jaguar extends Car {
// blah blah
}
#BoringCar
class Porche extends Car {
// blah blah
}
#BoringCar
class Ferrari extends Car {
// blah blah
}
#IncredibleCar
class Kia extends Car {
// blah blah
}
You could implement an AnnotationFilter class that removes cars from the list that do not have a certain annotation.
It might look something like this:
List<Car> carList = getListOfRandomCars();
AnnotationFilter<Car> annoFilter = new AnnotationFilter<Car>(BoringCar.class);
List<Car> boringCars = annoFilter.filter(carList);
Is that what you want to do?
If so, it can definitely be done.
The implementation for AnnotationFilter might look something like this:
public class AnnotationFilter<T> {
private Class filterAnno;
public AnnotationFilter(Class a) {
filterAnno = a;
}
public List<T> filter(List<T> inputList) {
if (inputList == null || inputList.isEmpty()) {
return inputList;
}
List<T> filteredList = new ArrayList<T>();
for (T someT : inputList) {
if (someT.getClass().isAnnotationPresent(filterAnno)) {
filteredList.add(someT);
}
}
return filteredList;
}
}
If that's not what you're after, a specific example would be helpful.
Java haven't got anything built in as such, which is why Reflections came about. Nothing built in that's as particular as what you're saying..
User-defined Annotations: we shall see how to annotate objects that we may come across in day-to-day life. Imagine that we want to persistent object information to a file. An Annotation called Persistable can be used for this purpose. An important thing is that we want to mention the file in which the information will get stored. We can have a property called fileName within the declaration of Annotation itself. The definition of the Persistable Annotation is given below,
Persistable.java
#Target({ElementType.FIELD, ElementType.LOCAL_VARIABLE})
public #interface Persistable
{
String fileName();
}
Annotations are just a way of tagging elements of a class; how these annotations are interpreted is up to the code that defines these annotations.
Is there some built-in way of handling annotations?
Annotations are used in so many different ways that it would be difficult to come up with a few "built-in ways" of handling them. There are source-level annotations (such as #Override and #Deprecated) that do not affect the behaviour of the code at all. Then there are runtime annotations that are usually very specific to a certain library, for eg. JAXB's binding annotations only make sense within a JAXBContext and Spring's autowiring annotations only make sense within an ApplicationContext. How would Java know what to do with these annotations simply by looking at a class which uses them?

Canonicalizing Java bean property names

I have a bunch of third-party Java classes that use different property names for what are essentially the same property:
public class Foo {
public String getReferenceID();
public void setReferenceID(String id);
public String getFilename();
public void setFilename(String fileName);
}
public class Bar {
public String getRefID();
public void setRefID(String id);
public String getFileName();
public void setFileName(String fileName);
}
I'd like to be able to address these in a canonicalized form, so that I can treat them polymorphically, and so that I can do stuff with Apache BeanUtils like:
PropertyUtils.copyProperties(object1,object2);
Clearly it would be trivial to write an Adapter for each class ...
public class CanonicalizedBar implements CanonicalizedBazBean {
public String getReferenceID() {
return this.delegate.getRefID();
}
// etc.
}
But I wonder is there something out there more generalized and dynamic? Something that would take a one-to-many map of property name equivalences, and a delegate class, and produce the Adapter?
I've never used it, but I think you're looking for Dozer:
Dozer is a Java Bean to Java Bean mapper that recursively copies data
from one object to another. Typically, these Java Beans will be of
different complex types.
Dozer supports simple property mapping, complex type mapping,
bi-directional mapping, implicit-explicit mapping, as well as
recursive mapping. This includes mapping collection attributes that
also need mapping at the element level.
Dozer not only supports mapping between attribute names, but also
automatically converting between types. Most conversion scenarios are
supported out of the box, but Dozer also allows you to specify custom
conversions via XML.
First Option is Dozer.
Second option is Smooks Framework
with a tweak. It will be beneficial to use Smook's Graphical mapper.
Another option would be XStream with custom Mapper.
maybe something like that:
public class CanonicalizedBar implements CanonicalizedBazBean {
public String getReferenceID() {
Method m = this.delegate.getClass().getDeclaredMethod("getReferenceID");
if(m == null)
m = this.delegate.getClass().getDeclaredMethod("getRefID");
...
return m.invoke();
}
// etc.
}
Although, I personally have never used it. I noticed that a project called orika is noted as having the best performance and the ability to automatically understand many such mappings.
At any rate it also supports custom mappings and uses generated code to implicitly define the adapters.
You can also define a custom mapper, that is if you know how to canonize the member names you can use that knowledge to build a mapping that is true for all your objects. for instance:
DefaultFieldMapper myDefaultMapper = new DefaultFieldMapper() {
public String suggestMapping(String propertyName, Type<?> fromPropertyType) {
// split word according to camel case (apache commons lang)
String[] words= StringUtils.splitByCharacterTypeCamelCase(propertyName);
if(words[0].length() > 6) {
// trim first camel-cased word of propery name to 3 letters
words[0]= words[0].substring(0,2);
return StringUtils.join(words);
} else {
// remains unchanged
return propertyName;
}
}
}
mapperFactory.registerDefaultFieldMapper(myDefaultMapper );
I haven't done much with it but you may be able to use Aspect Oriented Programming to do this.
What you should be able to do I think is add a method to each of the classes that internally calls the real method. See this article about half way down it talks about mixins.
AspectJ is probably the most popular implementation.

Magically call methods in Java

Is there some way of using magic methods in Java like there is in PHP with __call?
For instance:
class foo {
#Setter #Getter
int id;
#Getter
Map <String, ClassInFoo> myMap;
protected class ClassInFoo {
#Setter #Getter
String name;
}
#Setter
String defaultKey;
}
I'm using Project Lombok annotations for getter and setter methods to simplify the code.
Let's consider that that my map contains several items mapped by String and the defaultKey defines the default one.
What I would like is to be able to call foo.getName() which would return the default name as foo.myMap.get(defaultKey).getName().
The reason I can't just write all the getters manually is that the Foo class is in fact inherited with generics and the the inner class might be different.
I sort of need something like:
function Object __call(method) {
if (exist_method(this.method)
return this.method();
else
return this.myMap.get(defaultKey).method();
}
Is this somehow possible in Java?
EDIT:
I made a more precise example of what I am trying to achieve here: https://gist.github.com/1864457
The only reason of doing this is to "shorthand" the methods in the inner class.
You absolutely can through reflection by using its features like
public Method getMethod(String name, Class<?>... parameterTypes)
that can be used to see if a class has some methods defined but I don't see how your problem couldn't be solved with a proper use of interfaces, inheritance and overriding of methods
Features like reflection are provided to manage certain, otherwise unsolvable, issues but Java is not PHP so you should try to avoid using it when possible, since it's not in the philosophy of the language.
Isn't it the whole point of inheritance and overriding?
Base class:
public Object foo() {
return this.myMap.get(defaultKey).method();
}
Subclass:
#Overrides
public Object foo() {
return whateverIWant;
}

iBatis Discriminator on Insert

I have an abstract class Example and concrete subclasses to go along with it. I used a discriminator to pull data out of the database, like so:
<resultMap id="ExampleResultMap" class="Example">
<discriminator column="stateCode" javaType="java.lang.String">
<subMap value="AL" resultMap="AlabamaStateResultMap"/>
<subMap value="AR" resultMap="ArkansasStateResultMap"/>
[...]
</discriminator>
</resultMap>
<resultMap extends="ExampleResultMap"
id="AlabamaStateResultMap"
class="AlabamaState"/>
<resultMap extends="ExampleResultMap"
id="ArkansasStateResultMap"
class="ArkansasState"/>
[...]
Thus I have an AlabamaState object (a subclass of the abstract Example object) with no attributes of any kind on him. This is contrived, but the gist is that I don't have any attribute that uniquely identifies the object's type--and there's no reason I would if not for this case.
(Note: The classes aren't empty, they're behavioral, so refactoring them out of existence isn't an option.)
How do I save it back to the database?
Ideally there would be a Discriminator for ParameterMaps, but there doesn't seem to be one.
As far as I can tell, there are a number of undesirable solutions, among them:
Give up and add a "getType()" method on all my subclasses that returns a static string. In this case, AL. (Note that I tried pretty hard to avoid needing this throughout all my code, so having this = OOD-defeat).
Make a "DB" object that's exactly like my big, complex object but happens to also have an extra string saying "Oh, btw, my TYPE is AL."
Extract all 20 attributes I want to persist into a HashMap before inserting the object.
Some other craziness like using the toString() or something to help me out.
Likely I'll go with the first option, but it seems rather ridiculous, doesn't it? If iBatis can create it, shouldn't it be able to persist it? What I really need is a discriminator for insert.
Am I out of luck, or am I just overlooking something obvious?
If you have no attributes belonging to your subclasses, you should consider removing these subclasses and add an enum to your former base-class, since the only purpose your subclasses serve is to differentiate the type of your objects (if I understood you correctly). Using an enum for this is easier to extend and more elegant in client code (since you can switch on the enum instead of using blocks of instanceof expressions).
If are having special implementations of certain operations on your subclasses, you could move them to the enum as well, and have your base class delegate to the implementation on the enum.
EDIT
Here is an example:
public interface GreetingStrategy {
abstract String sayHello();
}
enum UserType implements GreetingStrategy {
ADMIN {
#Override
public String sayHello() {
return "hello from admin";
}
},
GUEST {
#Override
public String sayHello() {
return "hello from guest";
}
};
}
class User {
private final GreetingStrategy greetingStrategy;
public User(GreetingStrategy greetingStrategy) {
this.greetingStrategy = greetingStrategy;
}
public String sayHello() {
return greetingStrategy.sayHello();
}
}

Categories

Resources