The problem of generalized restructuring of databases has been addressed with two limitations: first, it is assumed that the restructuring user is able to describe the source and target databases in terms of the implicit data model of a particular methodology; second, the restructuring user is faced with the task of judging the scope and applicability of the defined types of restructuring to his database implementation and then of actually specifying his restructuring needs by translating them into the restructuring operations on a foreign data model. A certain amount of analysis of the logical and physical structure of databases must be performed, and the basic ingredients for such an analysis are developed here. The distinction between hierarchical and nonhierarchical data relationships is discussed, and a classification for database schemata is proposed. Examples are given to illustrate how these schemata arise in the conventional hierarchical and network systems. Application of the schema analysis methodology to restructuring specification is also discussed. An example is presented to illustrate the different implications of restructuring three seemingly identical database structures.