问题描述
由于历史原因,我需要将当前项目中使用EclipseLink的驼峰分隔名称映射到下划线名称。我知道我们可以在JPA中单独定制名称映射,但我们有一长串的驼峰名称要更改,所以我们想要尽可能避免使用那种样板代码。
I need to map camel-case names to underscore-separated names in my current project, which uses EclipseLink, due to historical reasons. I know we can customize name mapping individually in JPA, but we have a long list of camel-case names to change, so we want to avoid that kind of boilerplate codes if all possible.
我想要实现的目标如下。假设我们有一个实体类如下:
What I want to achieve is as follows. Suppose we have an entity class as below:
@Entity
public class FooBar {
@Id @GeneratedValue
private Long id;
@Temporal( TemporalType.TIMESTAMP )
private Date dateCreated;
}
我希望这个类映射到名为foo_bar的表和列 id和date_created。请注意,数据库中的所有名称都是小写的。
I want this class maps to a table with name "foo_bar" and columns "id" and "date_created". Note that all names in database are in lower case.
我搜索了一下,找到了更改表名的解决方案。但是,我无法弄清楚如何在实体类中更改字段名称。
I googled around, and found a solution for changing table names. However, I can't figure out how to change field names in an entity class.
下面是我的名字映射定制器,其中方法 updateFieldNameMappings()
未映射 fieldName
到 field_name
,这是我想要实现的。问题归结为如何在类定义中获取字段名称。那么,我如何在EclipseLink中做到这一点?
Below is my name-mapping customizer, where the method updateFieldNameMappings()
is not mapping fieldName
to field_name
, which is what I want to achieve. The problem boils down to how to get the field name as in the class definition. So, how do I do that in EclipseLink?
public class JpaNameMappingCustomizer implements SessionCustomizer {
@Override
public void customize( Session session ) throws Exception {
Map<Class, ClassDescriptor> descs = session.getDescriptors();
Collection<ClassDescriptor> descriptors = descs.values();
// This code assumes single table per descriptor!
for (ClassDescriptor desc : descriptors) {
updateTableNameMapping( desc );
updateFieldNameMapping( desc );
}
}
private void updateTableNameMapping ( ClassDescriptor desc ) {
Class clazz = desc.getJavaClass();
String tableName = camelCaseToUnderscore( clazz.getSimpleName() );
desc.setTableName( tableName );
}
private void updateFieldNameMapping ( ClassDescriptor desc ) {
// build name maps
Field[] fields = desc.getJavaClass().getDeclaredFields();
String tableName = desc.getTableName();
Map<String,String> nameMap = new HashMap<>();
String prefix = tableName + ".";
for( Field field : fields ) {
String name = field.getName();
String key = prefix + name.toUpperCase();
String value = prefix + camelCaseToUnderscore( name );
nameMap.put( key, value );
}
for (DatabaseMapping mapping : desc.getMappings()) {
if (mapping.isDirectToFieldMapping()) {
DirectToFieldMapping directMapping = (DirectToFieldMapping) mapping;
String oldFieldName = directMapping.getFieldName(); // format: table_name.FIELD
directMapping.setFieldName( nameMap.get( oldFieldName ) );
}
}
}
private String camelCaseToUnderscore( String camelCase ) {
return camelCase.trim().replaceAll("(?<!^)[A-Z](?!$)", "_$0").toLowerCase();
}
}
编辑11/10/13
我做了一些黑客攻击并改变了定制器。 updateFieldNameMapping()
仍然无法解决问题。在我看来,该方法中的语句 directMapping.setFieldName(nameMap.get(oldFieldName))
实际上并没有改变字段名称映射,这实在让我感到困惑。
edit 11/10/13I did some hacking and changed the customizer. The updateFieldNameMapping()
still doesn't do the trick. It seems to me that this statement directMapping.setFieldName( nameMap.get( oldFieldName ) )
inside the method doesn't actually change the field name mapping, which really confuses me.
编辑11/11/13
我忘了说清楚我有 eclipselink.session .customizer
在 persistence.xml
中启用。也就是说,我在 persistence.xml
中有如下所示的行:
edit 11/11/13I forgot to make it clear that I had the eclipselink.session.customizer
enabled in the persistence.xml
. Namely, I have a line like below in the persistence.xml
:
<property name="eclipselink.session.customizer" value="pkg.JpaNameMappingCustomizer"/>
推荐答案
而不是
directMapping.setFieldName(nameMap.get(oldFieldName ));
尝试:
directMapping.getField().resetQualifiedName(nameMap.get(oldFieldName));
这在EclipseLink 2.5.0下为我提供了诀窍
This does the trick for me under EclipseLink 2.5.0
这篇关于使用EclipseLink自定义JPA字段名称映射的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持!