Compare Non-Relational Database Systems With Relational Systems

Non-relational databases are simply lists of data, combining all records into one table without defining relationships between records or data items. Relational databases segregate data into logically organized tables, establishing how each table is linked, or related, to the database's other tables.

Data Organization

Non-relational databases such as flat file databases aggregate all their records in a single table. Relational databases break up data into smaller tables, eliminating redundancies in a process called normalization. Each table within a relational database has a unique identifier called a primary key which corresponds to another table's primary key, establishing a relationship between data within separate tables.

Data Integrity

Non-relational databases are prone to data duplication, threatening data integrity. Updating a person's address in a flat file database requires manually updating each of his records to ensure that his data remains consistent. The relational database design process eliminates record duplication, preventing inconsistent data from occupying the database.

Data Retrieval

Searches that use multiple criteria to find records in a non-relational database are inefficient, requiring multiple passes through the database to examine all its records for matches. Relational database organization allows a query language such as SQL to use the primary key shared between tables to quickly and efficiently collate and return the requested records.

References

• George Mason University: Healthcare Databases: Introduction to Evolution of Databases
• University of California, Santa Barbara: Linguistics UCCB: A Flat-File Database
• University of Notre Dame: What Are Relational Databases and Why Should I Care?

Resources

• Indiana University: University Information Technology Services: Knowledge Base: What is SQL?