You can combine mapping models sequentially. When you combine mapping models sequentially, you can merge two mapping models in such a way that mapping model A has the same target models as the source models in mapping model B. You use this combination when you want to calculate the direct mappings from a source schema to a target schema that you have already defined by an intermediate schema. For example:
The combined model only contains mappings that include TargetB in model A1 and SourceB in model A2. It might be possible that the combined mapping file is empty if there are no mappings that involve TargetB and SourceB.
You can combine the mapping models in a parallel order. You combine mapping models in parallel when you create multiple mapping models that might contain different sources but use the same target. You combine two of these mapping models to merge information into a third mapping model. The process of combining mapping models in parallel is especially useful when you have divergent teams that create mapping models, but you want all of the mappings to be contained in one mapping model. When the individual mapping models are complete, you can combine them to a single target. For example:
When you combine sources that are created on a federated server, the sources need to be defined to the same federated server. When mapping to a target model, queries are generated with the assumption that the user is deploying to the same server.
To combine mapping models, you must have at least one source and a target mapping model. The source and target mapping models that compose the new mapping model must be within the same project. Mappings to be combined must share either the source or the target. The source can be any physical database model. The target can be either any physical database model or any XML schema.