Object Database - Industry Disagreement

Industry Disagreement

There are two major groups of thought about what qualities are used to describe an OODBMS. The first is when the database supports object inheritance in a similar manner to an object-oriented programming language, and the second is that the database supports hierarchical data encapsulation in a similar way to XML, JSON or serialized PHP objects.

The first school of thought is, in reality, rare, as the underlying database design is inefficient, since it is often spread over several separate database tables. For example, if class 'Foo' extends class 'Bar', then the data defined within 'Foo' would be within one table and the data for 'Bar' would be within another. This separation is essential as class 'Goo' that also extends class 'Bar' will share the same 'Bar' table. This design then requires multiple reads, lock, logs and writes, and can slow all transactions down.

The second school of thought is often implemented within industry standard SQL databases by using standard text fields; however, this does have the effect of locking away the data held within as SQL only understands single or comma separated values. Object databases applying this school of thought will allow for data to be encoded in a hierarchical manner, while still allowing data to be indexed.