A method, apparatus, and computer instructions for defining queries in terms of data objects. A simplified query syntax is introduced for querying across a set of data objects which does not require the duplication of data object shape information. A data object is first defined by providing a name for the data object. Attribute and relationship information is added to the data object, wherein the information defines the shape of the data object. When a query is created, it is defined in terms of data objects. The query is processed, wherein processing the query includes extracting the shape of the data objects using the attribute information in the query. The query is expanded by adding additional syntax to the query based on relationship information defined for the data objects and then executed in its expanded form.