DATA CONVERSION RULES FROM NETWORK TO RELATIONAL DATABASES

Because of its popularity and user adaptability, many companies have chosen relational databases as their corporate database systems. However, with a huge investment in existing network and hierarchical database applications, there is a need for some companies to convert them into relational databases. Currently the simplest solution for data conversion is to write an application program for each network database file conversion or to develop a language to implement the conversion process. These are inefficient and costly. This paper describes a methodology for integrating schema translation and data conversion. Schema translation involves semantic reconstruction and the mapping of network schema into relational schema. Data conversion involves unloading records of each record type and their relationship relations into sequential files, and uploading them into relational table files. The conversion process can recover the semantics of the network schema by capturing users' knowledge. The methodology preserves the constraints of the network database by mapping the equivalent data dependencies of a loop-free network schema to a relational schema. The translation of navigational semantic from network schema to relational schema implies that each distinct access path of network schema should be considered as a subschema before conversion. The conversion process translates the existence and navigational semantics of the network database into a relational database without loss of information.