spring-data-jpa/src/docbkx/reference/jpa.xml

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE chapter PUBLIC "-//OASIS//DTD DocBook XML V4.4//EN"
"http://www.oasis-open.org/docbook/xml/4.4/docbookx.dtd">
<chapter id="jpa.repositories">
  <title>JPA Repositories</title>

  <abstract>
    <para>This chapter includes details of the JPA repository
    implementation.</para>
  </abstract>

  <section id="jpa.query-methods">
    <title>Query methods</title>

    <section id="jpa.sample-app.finders.strategies">
      <title>Query lookup strategies</title>

      <para>The JPA module supports defining a query manually as String or
      have it being derived from the method name.</para>

      <simplesect>
        <title>Declared queries</title>

        <para>Although getting a query derived from the method name is quite
        convenient one might face the situation in which either the method
        name parser does not support the keyword one wants to use or the
        method name would get unnecessarily ugly. So you can either use JPA
        named queries through a naming convention (see <xref
        linkend="jpa.query-methods.named-queries" /> for more information) or
        rather annotate your query method with
        <interfacename>@Query</interfacename> (see as for details).</para>
      </simplesect>

      <simplesect>
        <title>Strategies</title>

        <para>This strategy tries to find a declared query that can either be
        defined using JPA <code>@NamedQuery</code> means or Hades
        <code>@Query</code> annotation (see <xref
        linkend="jpa.query-methods.named-queries" /> and <xref
        linkend="jpa.query-methods.at-query" /> for details). If no declared
        query is found execution of the query will fail.</para>
      </simplesect>

      <simplesect>
        <title>CREATE_IF_NOT_FOUND (default)</title>

        <para>This strategy is actually a combination of the both mentioned
        above. It will try to lookup a declared query first but create a
        custom method name based query if no named query was found. This is
        default lookup strategy and thus will be used if you don't configure
        anything explicitly. It allows quick query definition by method names
        but also custom tuning of these queries by introducing declared
        queries for those who need explicit tuning.</para>
      </simplesect>
    </section>

    <section id="jpa.query-methods.query-creation">
      <title>Query creation</title>

      <para>Generally the query creation mechanism for JPA works as described
      in <xref linkend="repositories.query-methods" />. Here's a short example
      of what a JPA query method translates into:<example>
          <title>Query creation from method names</title>

          <para><programlisting language="java">public interface UserRepository extends Repository&lt;User, Long&gt; {

  List&lt;User&gt; findByEmailAddressAndLastname(String emailAddress, String lastname);
}</programlisting>We will create a query using the JPA criteria API from this
          but essentially this translates into the following query:</para>

          <programlisting>select u from User u where u.emailAddress = ?1 and u.lastname = ?2</programlisting>

          <para>Spring Data JPA will do a property check and traverse nested
          properties like described in <xref
          linkend="repositories.query-methods.property-expressions" />. Here's
          an overview of the keywords supported for JPA and what a method
          containing that keyword essentially translates to.</para>
        </example></para>

      <para><table>
          <title>Supported keywords inside method names</title>

          <tgroup cols="3">
            <colspec colwidth="1*" />

            <colspec colwidth="2*" />

            <colspec colwidth="3*" />

            <thead>
              <row>
                <entry>Keyword</entry>

                <entry>Sample</entry>

                <entry>JPQL snippet</entry>
              </row>
            </thead>

            <tbody>
              <row>
                <entry><code>And</code></entry>

                <entry><code>findByLastnameAndFirstname</code></entry>

                <entry><code>… where x.lastname = ?1 and x.firstname =
                ?2</code></entry>
              </row>

              <row>
                <entry><code>Or</code></entry>

                <entry><code>findByLastnameOrFirstname</code></entry>

                <entry><code>… where x.lastname = ?1 or x.firstname =
                ?2</code></entry>
              </row>

              <row>
                <entry><code>Between</code></entry>

                <entry><code>findByStartDateBetween</code></entry>

                <entry><code>… where x.startDate between 1? and
                ?2</code></entry>
              </row>

              <row>
                <entry><code>LessThan</code></entry>

                <entry><code>findByAgeLessThan</code></entry>

                <entry><code>… where x.age &lt; ?1</code></entry>
              </row>

              <row>
                <entry><code>GreaterThan</code></entry>

                <entry><code>findByAgeGreaterThan</code></entry>

                <entry><code>… where x.age &gt; ?1</code></entry>
              </row>

              <row>
                <entry><code>IsNull</code></entry>

                <entry><code>findByAgeIsNull</code></entry>

                <entry><code>… where x.age is null</code></entry>
              </row>

              <row>
                <entry><code>IsNotNull,NotNull</code></entry>

                <entry><code>findByAge(Is)NotNull</code></entry>

                <entry><code>… where x.age not null</code></entry>
              </row>

              <row>
                <entry><code>Like</code></entry>

                <entry><code>findByFirstnameLike</code></entry>

                <entry><code>… where x.firstname like ?1</code></entry>
              </row>

              <row>
                <entry><code>NotLike</code></entry>

                <entry><code>findByFirstnameNotLike</code></entry>

                <entry><code>… where x.firstname not like ?1</code></entry>
              </row>

              <row>
                <entry><code>OrderBy</code></entry>

                <entry><code>findByAgeOrderByLastnameDesc</code></entry>

                <entry><code>… where x.age = ?1 order by x.lastname
                desc</code></entry>
              </row>

              <row>
                <entry><code>Not</code></entry>

                <entry><code>findByLastnameNot</code></entry>

                <entry><code>… where x.lastname &lt;&gt; ?1</code></entry>
              </row>

              <row>
                <entry><code>In</code></entry>

                <entry><code>findByAgeIn(Collection&lt;Age&gt;
                ages)</code></entry>

                <entry><code>… where x.age in ?1</code></entry>
              </row>

              <row>
                <entry><code>NotIn</code></entry>

                <entry><code>findByAgeNotIn(Collection&lt;Age&gt;
                age)</code></entry>

                <entry><code>… where x.age not in ?1</code></entry>
              </row>
            </tbody>
          </tgroup>
        </table><note>
          <para><code>In</code> and <code>NotIn</code> also take any subclass
          of <interfacename>Collection</interfacename> as parameter as well as
          arrays or varargs.</para>
        </note></para>
    </section>

    <section id="jpa.query-methods.named-queries">
      <title>Using JPA NamedQueries</title>

      <note>
        <para>The examples use simple <code>&lt;named-query /&gt;</code>
        element and <code>@NamedQuery</code> annotation. The queries for these
        configuration elements have to be defined in JPA query language. Of
        course you can use <code>&lt;named-native-query /&gt;</code> or
        <code>@NamedNativeQuery</code>, too. These elements allow you to
        define the query in native SQL by losing the database platform
        independence.</para>
      </note>

      <simplesect>
        <title>XML named query definition</title>

        <para>To use XML configuration simply add the necessary
        <code>&lt;named-query /&gt;</code> element to the
        <filename>orm.xml</filename> JPA configuration file located in
        <filename>META-INF</filename> folder of your classpath. Automatic
        invocation of named queries is enabled by using some defined naming
        convention. For more details see below.</para>

        <example>
          <title>XML named query configuration</title>

          <programlisting language="xml">&lt;named-query name="User.findByLastname"&gt;
  &lt;query&gt;select u from User u where u.lastname = ?1&lt;/query&gt;
&lt;/named-query&gt;</programlisting>
        </example>

        <para>As you can see the query has a special name which will be used
        to resolve it at runtime.</para>
      </simplesect>

      <simplesect>
        <title>Annotation configuration</title>

        <para>Annotation configuration has the advantage not to need another
        config file to be edited, probably lowering maintenance cost. You pay
        for that benefit by the need to recompile your domain class for every
        new query declaration.</para>

        <example>
          <title>Annotation based named query configuration</title>

          <programlisting language="java">@Entity
@NamedQuery(name = "User.findByEmailAddress", 
  query = "select u from User u where u.emailAddress = ?1")
public class User {

}</programlisting>
        </example>
      </simplesect>

      <simplesect>
        <title>Declaring interfaces</title>

        <para>To allow execution of this named query all you need to do is to
        specify the <interfacename>UserRepository</interfacename> as
        follows:</para>

        <example>
          <title>Query method declaration in UserRepository</title>

          <programlisting language="java">public interface UserRepository extends JpaRepository&lt;User, Long&gt; {

  List&lt;User&gt; findByLastname(String lastname);

  User findByEmailAddress(String emailAddress);
}</programlisting>
        </example>

        <para>Declaring this method we will try to resolve a call to this
        method to a named query starting with the simple name of the
        configured domain class followed by the method name separated by a
        dot. So the example here would use the named queries defined above
        instead of trying to create a query from the method name.</para>
      </simplesect>
    </section>

    <section id="jpa.query-methods.at-query">
      <title>Using @Query</title>

      <para>Using named queries to declare queries for entities is a valid
      approach and works fine for a small number amount of queries. As the
      queries themselves are tied to a Java method to execute them you
      actually can bind them to the query executing methods using Spring Data
      JPA <code>@Query</code> annotation rather than annotating them to the
      domain class. This will free the domain class from persistence specific
      information and colocate the query to the repository interface.</para>

      <para>Querys annotated to the query method will trump queries defined
      using <code>@NamedQuery</code> or named queries declared in in
      <filename>orm.xml</filename>.</para>

      <example>
        <title>Declare query at the query method using @Query</title>

        <programlisting language="java">public interface UserRepository extends JpaRepository&lt;User, Long&gt; {

  @Query("select u from User u where u.emailAddress = ?1")
  User findByEmailAddress(String emailAddress);
}</programlisting>
      </example>
    </section>

    <section id="jpa.named-parameters">
      <title>Using named parameters</title>

      <para>By default Sprign Data JPA will use position based parameter
      binding as described in all the samples above. This makes query methods
      a little error prone to refactorings regarding the parameter position.
      To solve this issue you can use <code>@Param</code> annotation to give a
      method parameter a concrete name and bind the name in the query:</para>

      <example>
        <title>Using named parameters</title>

        <programlisting language="java">public interface UserRepository extends JpaRepository&lt;User, Long&gt; {

  @Query("select u from User u where u.firstname = :firstname or u.lastname = :lastname")
  User findByLastnameOrFirstname(@Param("lastname") String lastname,
                                 @Param("firstname") String firstname);
}</programlisting>

        <para>Note that the method parameters are switched according to the
        occurrence in the query defined.</para>
      </example>
    </section>

    <section id="jpa.modifying-queries">
      <title>Modifying queries</title>

      <para>All the sections before described how to declare queries to access
      a given entity or collection of entitites. Of course you can add custom
      modifying behaviour by using facilities described in <xref
      linkend="custom-implementations" />. As this approach is feasible for
      comprehensive custom functionality, you can achieve the execution of
      modifying queries that actually only need parameter binding by
      annotating the query method with <code>@Modifying</code>:</para>

      <example>
        <title>Declaring manipulating queries</title>

        <programlisting language="java">@Modifying
@Query("update User u set u.firstname = ?1 where u.lastname = ?2")
int setFixedFirstnameFor(String firstname, String lastname);</programlisting>
      </example>

      <para>This will trigger the query annotated to the method as updating
      query instead of a selecting one. As the
      <interfacename>EntityManager</interfacename> might contain outdated
      entities after the execution of the modifying query, we automatically
      clear it (see JavaDoc of
      <interfacename>EntityManager</interfacename>.<methodname>clear()</methodname>
      for details). This will effectively drop all non-flushed changes still
      pending in the <interfacename>EntityManager</interfacename>. If you
      don't wish the <interfacename>EntityManager</interfacename> to be
      cleared automatically you can set
      <interfacename>@Modifying</interfacename> annotation's
      <code>clearAutomatically</code> attribute to
      <literal>false</literal>;</para>
    </section>
  </section>

  <section id="specifications">
    <title>Specifications</title>

    <para>JPA 2 introduces a criteria API that can be used to build queries
    programatically. Writing a criteria you actually define the where-clause
    of a query for a query of the handled domain class. Taking another step
    back these criterias can be regarded as predicate over the entity that is
    verbalized by the JPA criteria API constraints.</para>

    <para>Spring Data JPA now takes the concept of a specification from Eric
    Evans' book Domain Driven Design, that carries the same semantics and
    provides an API to define such
    <interfacename>Specification</interfacename>s using the JPA criteria API.
    Thus, there is a <interfacename>JpaSpecificationExecutor</interfacename>
    interface you can additionally extend with your repository
    interface:</para>

    <programlisting language="java">public interface CustomerRepository extends CrudRepository&lt;Customer, Long&gt;, JpaSpecificationExecutor {
 …
}</programlisting>

    <para>The additional interface carries methods like this one, which allow
    you to execute <interfacename>Specification</interfacename>s in a variety
    of ways.</para>

    <programlisting language="java">List&lt;T&gt; readAll(Specification&lt;T&gt; spec);</programlisting>

    <para>The <interfacename>Specification</interfacename> interface now looks
    as follows:</para>

    <programlisting language="java">public interface Specification&lt;T&gt; {
  Predicate toPredicate(Root&lt;T&gt; root, CriteriaQuery&lt;?&gt; query,
            CriteriaBuilder builder);
}</programlisting>

    <para>Okay, so what is the typical use case?
    <interfacename>Specification</interfacename>s can easily be used to build
    an extensible set of predicates on top of an entity that then can be
    combined and used with <interfacename>JpaRepository</interfacename>
    without the need of declaring a query (method) for every needed
    combination of those. Here's an example:</para>

    <example>
      <title>Specifications for a Customer</title>

      <programlisting language="java">public class CustomerSpecs {

  public static Specification&lt;Customer&gt; isLongTermCustomer() {
    return new Specification&lt;Customer&gt;() {
      Predicate toPredicate(Root&lt;T&gt; root, CriteriaQuery&lt;?&gt; query,
            CriteriaBuilder builder) {

         LocalDate date = new LocalDate().minusYears(2);
         return builder.lessThan(root.get(Customer_.createdAt), date);
      }
    };
  }


  public static Specification&lt;Customer&gt; hasSalesOfMoreThan(MontaryAmount value) {
    return new Specification&lt;Customer&gt;() {
      Predicate toPredicate(Root&lt;T&gt; root, CriteriaQuery&lt;?&gt; query,
            CriteriaBuilder builder) {

         // build query here
      }
    };
  }
}</programlisting>
    </example>

    <para>Admittedly the amount of boilerplate leaves room for improvement
    (that will hopefully be reduced by Java 8 closures) but the client side
    becomes much nicer as you will see below. Besides that we have expressed
    some criteria on a business requirement abstraction level and created
    executable <interfacename>Specification</interfacename>s. So a client
    might use a <interfacename>Specification</interfacename> as
    follows:</para>

    <example>
      <title>Using a simple Specification</title>

      <programlisting language="java">List&lt;Customer&gt; customers = customerRepository.findAll(isLongTermCustomer());</programlisting>
    </example>

    <para>Okay, why not simply creating a query for this kind of data access?
    You're right. Using a single <interfacename>Specification</interfacename>
    does not gain a lot of benefit over a plain query declaration. The power
    of <interfacename>Specification</interfacename>s really shines when you
    combine them to create new <interfacename>Specification</interfacename>
    objects. You can achieve this through the
    <classname>Specifications</classname> helper class we provide to build
    expressions like this:</para>

    <example>
      <title>Combined Specifications</title>

      <para><programlisting language="java">MonetaryAmount amount = new MonetaryAmount(200.0, Currencies.DOLLAR);
List&lt;Customer&gt; customers = customerRepository.readAll(
  where(isLongTermCustomer()).or(hasSalesOfMoreThan(amount)));</programlisting>As
      you can see, <classname>Specifications</classname> offers some gluecode
      methods to chain and combine
      <interfacename>Specification</interfacename>s. Thus extending your data
      access layer is just a matter of creating new
      <interfacename>Specification</interfacename> implementations and
      combining them with ones already existing.</para>
    </example>
  </section>

  <section id="transactions">
    <title>Transactionality</title>

    <para>CRUD methods on repository instances are transactional by default.
    For reading operations the transaction configuration <code>readOnly</code>
    flag is set to true, all others are configured with a plain
    <classname>@Transactional</classname> so that default transaction
    configuration applies. For details see JavaDoc of
    <classname>Repository</classname>. If you need to tweak transaction
    configuration for one of the methods declared in
    <interfacename>Repository</interfacename> simply redeclare the method in
    your repository interface as follows:</para>

    <example>
      <title>Custom transaction configuration for CRUD</title>

      <programlisting language="java">public interface UserRepository extends JpaRepository&lt;User, Long&gt; {

  @Override
  @Transactional(timeout = 10)
  public List&lt;User&gt; findAll();

  // Further query method declarations
}</programlisting>

      <para>This will cause the <methodname>findAll()</methodname> method to
      be executed with a timeout of 10 seconds and without the
      <code>readOnly</code> flag.</para>
    </example>

    <para>Another possibility to alter transactional behaviour is using a
    facade or service implementation that typically covers more than one
    repository. Its purpose is to define transactional boundaries for non-CRUD
    operations:</para>

    <example>
      <title>Using a facade to define transactions for multiple repository
      calls</title>

      <programlisting language="java">@Service
class UserManagementImpl implements UserManagement {

  private final UserRepository userRepository;
  private final RoleRepository roleRepository;

  @Autowired
  public UserManagementImpl(UserRepository userRepository, 
    RoleRepository roleRepository) {
    this.userRepository = userRepository;
    this.roleRepository = roleRepository;
  }

  @Transactional
  public void addRoleToAllUsers(String roleName) {

    Role role = roleRepository.findByName(roleName);

    for (User user : userRepository.readAll()) {
      user.addRole(role);
      userRepository.save(user);
    }
}</programlisting>

      <para>This will cause call to
      <methodname>addRoleToAllUsers(…)</methodname> to run inside a
      transaction (participating in an existing one or create a new one if
      none already running). The transaction configuration at the repositories
      will be neglected then as the outer transaction configuration determines
      the actual one used. Note that you will have to activate
      <code>&lt;tx:annotation-driven /&gt;</code> explicitly to get annotation
      based configuration at facades working. The example above assumes you're
      using component scanning.</para>
    </example>

    <section id="transactional-query-methods">
      <title>Transactional query methods</title>

      <para>To let your query methods be transactional simply use
      <interfacename>@Transactional</interfacename> at the repository
      interface you define.</para>

      <example>
        <title>Using @Transactional at query methods</title>

        <programlisting language="java">@Transactional(readOnly = true)
public interface UserRepository extends JpaRepository&lt;User, Long&gt; {

  List&lt;User&gt; findByLastname(String lastname);

  @Modifying
  @Transactional
  @Query("delete from User u where u.active = false")
  void deleteInactiveUsers();
}</programlisting>

        <para>Typically you will use the <code>readOnly</code> flag set to
        true as most of the query methods will be reading ones. In contrast to
        that <methodname>deleteInactiveUsers()</methodname> makes use of the
        <interfacename>@Modifying</interfacename> annotation and overrides the
        transaction configuration. Thus the method will be executed with
        <code>readOnly</code> flag set to false.</para>
      </example>

      <note>
        <para>It's definitely reasonable to use transactions for read only
        queries as we can mark them as such by setting the
        <code>readOnly</code> flag. This will not act as check that you do not
        trigger a manipulating query nevertheless (although some databases
        reject e.g. <literal>INSERT</literal> or <literal>UPDATE</literal>
        statements inside a transaction set to be read only) but gets
        propagated as hint to the underlying JDBC driver to do performance
        optimizations. Furthermore Spring will do some optimizations to the
        underlying JPA provider. E.g. when used with Hibernate the flush mode
        is set to <code>NEVER</code> when you configure a transaction as read
        only which causes Hibernate to skip dirty checks that gets quite
        noticeable on large object trees.</para>
      </note>
    </section>
  </section>

  <section id="jpa.auditing">
    <title>Auditing</title>

    <para>Most applications will require some auditability for entities
    allowing to track creation date and user and modification date and user.
    Spring Data JPA provides facilities to add this audition information to
    entity transparently by AOP means. To take part in this functionality your
    domain classes have to implement a more advanced interface:</para>

    <example>
      <title><interfacename>Auditable</interfacename> interface</title>

      <programlisting language="java">public interface Auditable&lt;U, ID extends Serializable&gt;
        extends Persistable&lt;ID&gt; {

    U getCreatedBy();

    void setCreatedBy(U createdBy);

    DateTime getCreatedDate();

    void setCreated(Date creationDate);

    U getLastModifiedBy();

    void setLastModifiedBy(U lastModifiedBy);

    DateTime getLastModifiedDate();

    void setLastModified(Date lastModifiedDate);
}</programlisting>
    </example>

    <para>As you can see the modifying entity itself only has to be an entity.
    Mostly this will be some sort of User entity, so we chose U as parameter
    type.</para>

    <note>
      <para>To minimize boilerplate code Spring Data JPA offers
      <classname>AbstractPersistable</classname> and
      <classname>AbstractAuditable</classname> base classes that implement and
      preconfigure entities. Thus you can decide to only implement the
      interface or enjoy more sophisticated support by extending the base
      class.</para>
    </note>

    <simplesect>
      <title>General auditing configuration</title>

      <para>Spring Data JPA ships with an entity listener that can be used to
      trigger capturing auditing information. So first you have to register
      the <classname>AuditingEntityListener</classname> inside your
      <filename>orm.xml</filename> to be used for all entities in your
      persistence contexts:</para>

      <example>
        <title>Auditing configuration orm.xml</title>

        <programlisting language="xml">&lt;persistence-unit-metadata&gt;
  &lt;persistence-unit-defaults&gt;
    &lt;entity-listeners&gt;
      &lt;entity-listener class="….data.jpa.domain.support.AuditingEntityListener" /&gt;
    &lt;/entity-listeners&gt;
  &lt;/persistence-unit-defaults&gt;
&lt;/persistence-unit-metadata&gt;</programlisting>
      </example>

      <para>Now activating auditing functionlity is just a matter of adding
      the Spring Data JPA <literal>auditing</literal> namespace element to
      your configuration:</para>

      <example>
        <title>Activating auditing in the Spring configuration</title>

        <programlisting language="xml">&lt;jpa:auditing auditor-aware-ref="yourAuditorAwareBean" /&gt;</programlisting>
      </example>

      <para>As you can see you have to provide a bean that implements the
      <interfacename>AuditorAware</interfacename> interface which looks as
      follows:</para>

      <example>
        <title><interfacename>AuditorAware</interfacename> interface</title>

        <programlisting language="java">public interface AuditorAware&lt;T, ID extends Serializable&gt; {

    T getCurrentAuditor();
}</programlisting>
      </example>

      <para>Usually you will have some kind of authentication component in
      your application that tracks the user currently working with the system.
      This component should be <interfacename>AuditorAware</interfacename> and
      thus allow seemless tracking of the auditor.</para>
    </simplesect>
  </section>

  <section id="jpa.misc">
    <title>Miscellaneous</title>

    <section>
      <title>Merging persistence units</title>

      <para>Spring supports ahving multiple persistence units out of the box.
      But sometimes you might wanna modularize your application but make sure
      that all these module run inside a single persistence unit at runtime.
      To do so Spring Data JPA offers a PersistenceUnitManager implementation
      that automatically merges persistence units based on their name.</para>

      <example>
        <title>Using MergingPersistenceUnitmanager</title>

        <para><programlisting language="xml">&lt;bean class="….LocalContainerEntityManagerFactoryBean"&gt;
  &lt;property name="persistenceUnitManager"&gt;
    &lt;bean class="….MergingPersistenceUnitManager" /&gt;
  &lt;/property
&lt;/bean&gt;</programlisting></para>
      </example>
    </section>
  </section>
</chapter>