Why another tool to dump the JPA schema? All tools that we've found were related to legacy versions of Hibernate or were covering just simple cases, without options to configure dialect or naming strategy. Also all of the tools we've found are based on the SchemaExport
class, which does not always correlate with the runtime schema - for example due to the lack of support for the Integrator
services, used to register UserType
classes like JodaTime or similar. We were also looking for a tool that is be able to handle further schema migrations, not just dump the current version.
packages
(required): list of packages containing JPA entitiesoutputPath
: output file for the generated schema. By default:- for
UPDATE
action:BUILD_OUTPUT_DIR/generated-resources/scripts/
- for other actions:
BUILD_OUTPUT_DIR/generated-resources/scripts/database.sql
jpaProperties
: additional properties like dialect or naming strategies which should be used in generation task. By defaultempty
formatOutput
: should the output be formatted. By defaulttrue
skipSequences
: should the generator skip sequences creation. By defaultfalse
delimiter
: delimiter used to separate statements. By default;
action
: which statements should be generated. By default:CREATE
. Possible values:DROP
CREATE
DROP_AND_CREATE
UPDATE
generationMode
: schema generation mode. By defaultDATABASE
. Possible values:DATABASE
: generation based on setting up embedded database and dumping the schemaMETADATA
: generation based on static metadata
processorProperties
: properties passed to externalSchemaProcessor
classes
H2 is often used to imitate native database engines, however with usage going beyond simple SQL it has huge limitations. To resolve some of them related to sequences we provide custom database dialects.
com.devskiller.jpa2ddl.dialects.H2PostgreSQL95Dialect
: Postgres 9.5 dialect for H2com.devskiller.jpa2ddl.dialects.H2PostgreSQL10Dialect
: Postgres 10 dialect for H2com.devskiller.jpa2ddl.dialects.H2MySQL57Dialect
: MySQL 5.7 dialect for H2com.devskiller.jpa2ddl.dialects.H2MySQL80Dialect
: MySQL 8.0 dialect for H2
If you need different dialects please build them using above examples.
You can run this plugin directly or integrate it into the default build lifecycle.
<build>
<plugins>
<plugin>
<groupId>com.devskiller.jpa2ddl</groupId>
<artifactId>jpa2ddl-maven-plugin</artifactId>
<version>0.9.12</version>
<extensions>true</extensions> <!-- required to run automatically -->
<configuration>
<packages>
<package>com.test.model</package>
</packages>
</configuration>
</plugin>
</plugins>
</build>
<build>
<plugins>
<plugin>
<groupId>com.devskiller.jpa2ddl</groupId>
<artifactId>jpa2ddl-maven-plugin</artifactId>
<version>0.9.12</version>
<extensions>true</extensions> <!-- required to run automatically -->
<configuration>
<outputPath>${basedir}/src/main/resources/database.sql</outputPath>
<packages>
<package>com.test.model</package>
<package>com.test.entities</package>
</packages>
<jpaProperties>
<property>
<name>hibernate.dialect</name>
<value>org.hibernate.dialect.MySQL57Dialect</value>
</property>
<property>
<name>hibernate.default_schema</name>
<value>prod</value>
</property>
</jpaProperties>
<formatOutput>true</formatOutput>
<skipSequences>true</skipSequences>
<delimiter>;</delimiter>
<action>DROP_AND_CREATE</action>
</configuration>
</plugin>
</plugins>
</build>
It's also possible to generate automated migrations scripts. JPA2DDL supports Flyway naming patterns for versioned migrations.
All subsequent migration scripts are saved in the outputPath
, in the following layout:
src/main/resources/migrations/
v1__jpa2ddl.sql
v2__jpa2ddl.sql
... next
Please note that after generation you can change the name of the file to make it more descriptive following the filename pattern v(N)__jpa2ddl(_custom_description).sql
- for example v1__jpa2ddl_init.sql
Sample configuration:
<build>
<plugins>
<plugin>
<groupId>com.devskiller.jpa2ddl</groupId>
<artifactId>jpa2ddl-maven-plugin</artifactId>
<version>0.9.12</version>
<extensions>true</extensions> <!-- required to run automatically -->
<configuration>
<outputPath>${basedir}/src/main/resources/migrations/</outputPath>
<packages>
<package>com.test.model</package>
<package>com.test.entities</package>
</packages>
<jpaProperties>
<property>
<name>hibernate.dialect</name>
<value>org.hibernate.dialect.MySQL57Dialect</value>
</property>
</jpaProperties>
<formatOutput>true</formatOutput>
<delimiter>;</delimiter>
<action>UPDATE</action>
</configuration>
</plugin>
</plugins>
</build>
./mvnw com.devskiller.jpa2ddl:jpa2ddl-maven-plugin:0.9.12:generate
Below you can find a sample build.gradle
script configuration:
buildscript {
repositories {
mavenCentral()
}
dependencies {
classpath "com.devskiller.jpa2ddl:jpa2ddl-gradle-plugin:0.9.12"
}
}
apply plugin: 'com.devskiller.jpa2ddl'
jpa2ddl {
packages = ['com.test.model']
}
Sometimes more actions that just saving the database migrations are needed.
Example of such use case is when there is a need to generate QueryDSL or JOOQ mappings.
The SchemaProcessor
mechanism in jpa2ddl resolves such needs.
Additional dependency jpa2ddl-querydsl-processor
provides the processor to generate mappings for QueryDSL.
To enable it:
- add a
jpa2ddl-querydsl-processor
dependency to the plugin (plugin->dependencies->dependency
) - configure the processor in the
plugin->configiration->processorProperties
section:queryDslOutputPath
: output path for generated mapping classesqueryDslOutputPackage
: optionally set package for generated classes
<build>
<plugins>
<plugin>
<groupId>com.devskiller.jpa2ddl</groupId>
<artifactId>jpa2ddl-maven-plugin</artifactId>
<version>${project.version}</version>
<extensions>true</extensions>
<configuration>
<packages>
<package>oss.devskiller.model</package>
</packages>
<action>UPDATE</action>
<processorProperties>
<property>
<name>queryDslOutputPath</name>
<value>${project.build.directory}/generated-sources/query-dsl</value>
</property>
<property>
<name>queryDslOutputPackage</name>
<value>oss.devskiller.querydsl</value>
</property>
</processorProperties>
</configuration>
<dependencies>
<dependency>
<groupId>com.devskiller.jpa2ddl</groupId>
<artifactId>jpa2ddl-querydsl-processor</artifactId>
<version>${project.version}</version>
</dependency>
</dependencies>
</plugin>
</plugins>
</build>
It's possible to build and inject your custom schema processors to jpa2ddl.
The only thing you need to do is to implement the com.devskiller.jpa2ddl.SchemaProcessor
interface, and add the jar with our implementation as a dependency for the plugin.
Please refer to the jpa2ddl-querydsl-processor to see an example.