MyBatis association collection and discriminator

2016/04/17 MyBatis Mybatis, Java

Quick note about MyBatis association, collection and discriminator.

association

<association> deals with a “has-one” type relationship. For example, in our example, an Author has an User account with username and password:

public class User {

	private int id;
	private String userName;
	private String password;

	// Getters and Setters
}

public class Author {

	private Integer id;
	private String realName;
	private String IDCard;
	private User user; // In DB, author table has a foreign key userID refering id in user table
	
	// Getters and Setters
}

And in xml map file, we need to define result Map of `Author`:

<resultMap id="AuthorMap" type="Author">
	<!-- author.id is tableName.id which is used to be distinguished from user.id -->
	<id property="id" column="author.id"/>
	<result property="realName" column="realName" />
	<result property="IDCard" column="IDCard" />
	<association property="user" column="userID" javaType="User">
		<id property="id" column="user.id"/>
		<result property="userName" column="userName" />
		<result property="password" column="password" />
	</association>
</resultMap>
<select id="selectAuthorJoin" resultMap="AuthorMap">
	select * from author inner join user on user.id = author.userID
</select>

We could also extract User information as another individual resultMap which could be re-used:

<resultMap id="AuthorMap" type="Author">
	...
	<association property="user" column="userID" javaType="User" resultMap="userMap"/>
</resultMap>
<resultMap id="userMap" type="User">
	<id property="id" column="id"/>
	<result property="userName" column="userName" />
	<result property="password" column="password" />
</resultMap>

To query by java:

List<User> userList = session.selectList("selectAuthorJoin");

Constructor

If we do not care about user id or some other property. We could use <constructor> to limit what we want based on constructors of object.

public User() {}

public User(String userName, String password) {
	this.userName = userName;
	this.password = password;
}
<resultMap id="AuthorMapByConstructor" type="Author">
	<id property="id" column="author.id" />
	<result property="realName" column="realName" />
	<result property="IDCard" column="IDCard" />
	<association property="user" column="userID" javaType="JiKeUser">
		<constructor>
			<arg column="userName" javaType="String" />
			<arg column="password" javaType="String" />
		</constructor>
	</association>
</resultMap>
<select id="selectAuthorJoin" resultMap="AuthorMapByConstructor">
	select * from author inner join user on user.id = author.userID
</select>

Subquery

Subquery/Inner query/Nested query is a query within another SQL query and embedded within the WHERE clause.

select * from author where userID in (select id from user)

So there are in fact two statements here. Let’s define each select statement and resultMap:

<resultMap id="AuthorSubMap" type="Author">
	<id property="id" column="author.id" />
	<result property="realName" column="realName" />
	<result property="IDCard" column="IDCard" />
	<association property="user" column="userID" javaType="User" select="findById"/>
</resultMap>
<select id="findById" parameterType="int" resultType="User">
	select * from user where id=#{id}
</select>
<select id="selectAuthorSub" resultMap="AuthorSubMap">
	select * from author 
</select>

In the above settings, <association> will pass userID as parameter to findById.

N+1 problem

But think about we have 10 authors, so here findById will be called 10 times! So we need to call 10 + 1 statements. It’s called “N+1” problem of subquery. Many people think it’s not as good as join query.

Solution

The solution is lazy load. Statement will be not queried until the point at which it is needed.

Let’s firstly find differences between subquery and join query.

  • Join query: Query one time. But it gets all the properties of two table, so it consumes resources.
  • Subquery: Query N+1 times. It depends on what we want in the final result. With lazy loading, we may not need to make N times queries if not necessary.

To set lazy loading in configuration file, this setting must be in front of others:

<settings>
	<setting name="lazyLoadingEnabled" value="true"/>
	<setting name="aggressiveLazyLoading" value="false"/>
</settings>

To test:

List<Author> authorList = session.selectList("selectAuthorSub");
for(Author author:authorList) {

	System.out.println(auther.getRealName());
	System.out.println("Lazyloading");
	// findById query will not be called until getUser() is invoked. It's lazy loading here.
	// If we never call getUser(), so findById will also never be called
	System.out.println(auther.getUser().getUserName());
	
}

So here with lazy loading, if we never need to call getUser(), it only queries one time which is efficient!

collection

<collection> element works almost identically to <association>. But it is used to map a set of nested results like List.

In the following example, each User could have a list of visitors:

public class User {

	private int id;
	private String userName;
	private String password;
	private List<Visitor> visitorList;

	// Getters and Setters
}

public class Visitor {

    private Integer visitID;
    private Date visitDate;
    private String visitIP;
    private User user; 

    // Getters and Setters
}

Then in xml map file, we use <collection> to indicate that list of visits.

<resultMap id="visitMap" type="User">
	<id property="id" column="id" />
	<result property="userName" column="userName" />
	<!-- `ofType` indicates the containing type of ArrayList -->
	<collection property="visitorList" javaType="ArrayList" column="visitID" ofType="Visitor">
		<result property="visitID" column="visitID" />
		<result property="visitIP" column="visitIP" />
		<result property="visitDate" column="visitDate" />
	</collection>
</resultMap>

<select id="selectVisit" resultMap="visitMap">
	select * from user inner join visitor on user.id = visit.userID
</select>

To test:

List<User> userList = session.selectList("selectVisit"); 

for(User user:userList) {

	System.out.println(user.getUserName());

	for(Visitor visitor:user.getVisitorList()) {

		System.out.println(visitor.getVisitDate() + visitor.getVisitIP());

	}	
}

discriminator

A single database query might return result sets of many different data types. So here <discriminator> is used to determine which data types or so-called resultMap to use according to a column value. It is like switch in other programming languages.

The following resultMap will return different data types according to value of vehicle_type.

<resultMap id="vehicleResult" type="Vehicle">
  <id property="id" column="id" />
  ...
  <discriminator javaType="int" column="vehicle_type">
    <case value="1" resultType="carResult">
      <result property="doorCount" column="door_count" />
    </case>
    <case value="2" resultType="truckResult">
      <result property="boxSize" column="box_size" />
      <result property="extendedCab" column="extended_cab" />
    </case>
    <case value="3" resultType="suvResult">
      <result property="allWheelDrive" column="all_wheel_drive" />
    </case>
  </discriminator>
</resultMap>

Refs

Search

    Directory