You can use the <attribute> tag as a child tag in a <property> or <item-descriptor> tag to associate arbitrary name/string value pairs with the property. These named values correspond to those specified in java.beans.FeatureDescriptor, which RepositoryPropertyDescriptor implements.

This is simply a way of letting applications associate more metadata with individual properties. The SQL repository does not do anything with the data expressed in the <attribute> tag; it remembers the values defined in the template and allows one to read them at runtime. This is the same mechanism that the @beaninfo Javadoc tag system uses.

Here is an example, which assigns to the author property the name/value pair of maxLength="30":

<!-- The "book" item type -->
<item-descriptor name="book">
  <table name="book" id-column-names="book_id" type="primary">
    <property name="title"/>
    <property name="author">
      <attribute name="maxLength" value="30"/>
    </property>
  </table>
</item-descriptor>

It is also useful to be able to refer to values of Nucleus components as attributes. You can do this with the bean attribute of the <attribute> tag. For example:

<attribute name="documentRootPath"
     bean="/atg/demo/QuincyFunds/repositories/
FeaturesDataStore.relativePathPrefix" />

If you use a relative Nucleus address for the bean attribute, it refers to a component relative to the Repository component.

You can access property attributes programmatically using the RepositoryPropertyDescriptor.getValue method. For example:

RepositoryPropertyDescriptor.getValue("maxLength");
 
loading table of contents...