Method for generating SQL commands to create an integrated global schema

• US 5,819,277 A
• Filed: 03/26/1996
• Issued: 10/06/1998
• Est. Priority Date: 03/26/1996
• Status: Expired due to Term

• Current Assignee: Bull HN Information Systems Incorporated (Atos SE)
• Original Assignee: Bull HN Information Systems Incorporated (Atos SE)
• Inventors: Golshani, Forouzan, Howell, Thomas H., Friesen, Oris D.
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): Min, Donald
• Application Number:
  US08/624,723
• Time in Patent Office: 924 Days
• Field of Search: 707/4, 707/100, 707/101, 707/102, 707/103, 707/10
• US Class Current: 1/1
• CPC Class Codes: G06F 16/289 Object oriented databases
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation

Abstract

A method practiced by a computer system for generating SQL commands to create an integrated global schema. The input is two relational schemas in object-oriented form and a list of correspondences between components, object classes, objects and attributes. Based on correspondences object classes within one schema are folded into one object class, or relation, and likewise, based on correspondences one relation from each input schema is merged into a single relation of the integrated schema. This process is continued until no more object classes can be folded together and no more can be merged. SQL commands are generated based on specified correspondences between attributes depending upon whether a relation of the integrated relational schema was derived from only one relation of one of the input schemas, or is the result of the merger of two relations each from a different input schema. In the latter case, correspondences are based on the attributes of merged relations.

Claims

1. A method residing in a computer system for integrating two relational schemas in object-oriented form to produce an integrated schema and for generating SQL commands for registering relations of the integrated schema:

comprising the steps of;

a, receiving as inputs two relational input schemas of databases to be integrated in object-oriented form and specified correspondences between components, object classes, objects and attributes of said two schemas;

b, folding together object classes within the same schema into one schema based on respective correspondences of attributes;

c, merging two object classes, each object class being from a different input schema, based on assertions with respect to the attributes of said two object classes to produce an integrated object class of the integrated schema;

d, repeating steps b and c until there are no more correspondences between attributes of the object classes of either of said input schemas that can be folded together and no more object classes that can be merged;

e, creating an object class in the integrated schema for every object class of the two input schemas that remain after the completion of step d;

f, generating SQL commands based on specified correspondences between attributes of the relations, with each SQL command generating a relation of the integrated schema depending upon whether the integrated relation corresponds to one and only one relation of one of the input schemas, or is the result of the merger of two relations one from each of the input schemas, and if the latter on specified correspondences of the attributes of the two relations of the input schemas.

2. The method of claim 1 in which in step f, the integrated relation is the result of the merger of two relations one from each of the input schemas based on correspondences between the primary key attributes of said two relations.

3. The method of claim 2 in which in step f, the integrated relation is the result of the merger of two relations one from each of the input schemas based on correspondences between the primary key attributes and other attributes of said two relations.

4. The method of claim 2 in which in step f, the integrated relation is the result of the merger of two relations one from each of the input schemas based on correspondences between non-primary key attributes of said two relations.

5. A method residing in a computer system for integrating two relational schemas in object-oriented form and for generating SQL commands for registering relations of the integrated database;

comprising the steps of;

a, receiving as inputs two relational input schemas of databases to be integrated in object-oriented form and a specified list of correspondences between components, relations and attributes, of said two schemas;

b, selecting a pair of relations that correspond from the correspondence list;

c, folding together corresponding relations within the same schema;

d, merging two corresponding relations one from each of the two input schemas to form a new integrated relation;

e, creating a list of attributes for the newly integrated relation including all of the attributes of the two relations from which formed;

f, repeating steps b-e until there are no more relations that correspond;

g, creating a relation for every relation of the two input schemas remaining at the completion of step f in the integrated schema;

h, generating SQL commands for each relation of the integrated schema based on whether derived from a relation corresponding to one and only one input schema, or on whether derived from two relations, one from each of the input schemas; and

if derived from two relations, then based on specified correspondence between attributes of said two relations.

6. The method of claim 5 in which in step h, each relation of the integrated schema derived from two relations, one from each of the input schemas is based on correspondences between the primary key attributes of said two relations.

7. The method of claim 5 in which in step h, each relation of the integrated schema derived from two relations, one from each of the input schemas is based on correspondences between the primary key attributes and other attributes of said two relations.

8. The method of claim 5 in which in step h, each relation of the integrated schema derived from two relations, one from each of the input schemas is based on correspondences between the non-primary key attributes of said two relations.