DAO with Null Object Pattern

Question

After Reading:
Effective Java (See Item 43) - Joshua Bloch
Clean Code (Don't Return Null) - Uncle Bob
Avoiding != null statements
Null Object pattern

I was looking for an answer to the question of what a DAO should return when a search ends up to be for an entity that does not exist for non-collection objects. Collection object is really ok by using empty array or emptyList methods. But with non-collections it might be harder. An alternative solution is to never return null and instead use the Null Object pattern. But I have no idea to integrate with Null Object pattern with DAO and I really excited to see great integration with Null Object pattern and DAO pattern especially for model(dto) object return case.

I would appreciate and welcome any best design pattern, scenario and suggestion.

Answer 1

Indeed introducing null reference is probably one of the worse mistake in the programming languages' history even its creator Tony Hoare calls it his billion-dollar mistake.

Here are the best alternatives to null according to your Java version:

1. Java 8 and above

Starting from Java 8 you can use java.util.Optional.

Here is an example of how you could use it in your case:

public Optional<MyEntity> findMyEntity() {
    MyEntity entity = // some query here
    return Optional.ofNullable(entity);
}

2. Prior to Java 8

Before Java 8 you can use com.google.common.base.Optional from Google Guava.

Here is an example of how you could use it in your case:

public Optional<MyEntity> findMyEntity() {
    MyEntity entity = // some query here
    return Optional.fromNullable(entity);
}

Answer 2

All you need to do is return an empty object - say a customer entry you would have in your DAO something like

if (result == null) { return new EmptyUser(); }

where EmptyUser extends User and returns appropriate entries to getter calls to allow the rest of your code to know it is an empty object (id = -1 etc)

A small example

public class User {
    private int id;
    private String name;
    private String gender;
    public String getName() {
       //Code here
    }
    public void setName() {
       //Code here
    }
}

public class EmptyUser extends User {

    public int getId() {
       return -1;
    }

    public String getName() {
       return String.Empty();
   }
}

public User getEntry() {
   User result = db.query("select from users where id = 1");
   if(result == null) {
       return new EmptyUser();
   }
   else {
       return result;
    }
}

Answer 3

In my experience, in real-life scenarios where a single entity should be returned returning a null is actually either a data inconsistency error or lack of data whasoever. In both cases the really good thing to do is to crate and throw your own DataNotFoudException. Fail fast.

I'm using mybatis as ORM and recently I started writing some theoretically single-result mapper selects as returning lists and validating checking the amount of returned data in dao, and throwing exceptions when the returned amounts do not match dao's method assumptions. It works pretty well.