Developing with the SDK > Using and Writing Data Models > Deciding Your Data Model Strategy

First you should decide which classes in the model package are appropriate, depending on the characteristics of your existing business data and your requirements.

There are five levels of data model implementation, each appropriate for certain situations:

  1. No use of existing data.
  2. There are three possible situations:
    In these cases, use the default SDM model implementation, ilog.views.sdm.model.IlvDefaultSDMModel.
    The nodes and the links of the default model are instances of the classes IlvDefaultSDMNode and IlvDefaultSDMLink. . These objects have an arbitrary set of user-defined properties, so the default model can represent all kinds of objects.
  3. Display existing data in a database.
  4. Your data is stored in a database, and you want to display it but not modify it directly in the diagram component.
    The model package provides a specific SDM model for this case: ilog.views.sdm.model.IlvJDBCSDMModel. This model uses the JDBC API to access the database. All you need to do is configure the model with the database URL, the names of the tables containing the nodes and links, and a few other parameters.
  5. Existing JavaBeans.
  6. You have existing data stored in classes which conform to the JavaBeans standard: there are accessor methods (get and set) for each Bean property.
    In this case, use the class ilog.views.sdm.model.IlvJavaBeanSDMModel. This model uses your JavaBeans as the nodes and links of a graph, so there is no duplication or additional memory consumption. The properties of nodes and links are read and written using the JavaBeans API. All you have to do is supply the set of JavaBeans that will make up your graph. For an example, see JavaBeans Example.
  7. Existing Java classes that are not JavaBeans or existing data but no Java classes.
  8. For example, you already have an implementation of a Swing JTree model that represents your data and you want to display the contents of the tree without duplicating data.
    One of the following applies:
    In such cases, there are two possible approaches:

Table 2.1 summarizes the strategies described in this section.

Table 2.1 Summary of Data Model Strategies
Data Characteristics 
Recommended Class 
New or from XML 
Database (display only) 
JavaBeans 
Java classes or other existing objects (for example, Swing) 

If the graphic objects are to be moved interactively, SDM tries to store the new object locations as x,y properties in the model. However, the model can refuse to set the x,y properties, for example when it is a read-only model. To cope with such a case, the SDM library provides a metadata system that manages new properties as if they were stored in the true model.

To enable metadata, call ilog.views.sdm.IlvSDMEngine.setMetadataEnabled(true). Metadata is saved only in XML files. Other persistant mechanisms are responsible for saving metadata if they need to.

The default model does not need metadata.

images/dia_sdm_model.png

Figure 2.1 The SDM Model Class Relationships

images/dia_abstract_sdm_model.png

Figure 2.2 The Abstract SDM Model Class Relationships