Data modeling is an important part of application development so you understand how to structure the data. Decisions are made based on the application requirements, what information is needed, and how it is needed.

We could probably discuss for another couple chapters the value of normalization versus denormalization of the data model and whether to use surrogate or natural keys on tables, but those topics are not necessarily the point of the data management and model discussion here.

During application development, there are decisions being made to use data in a particular way based on the requirements. This might even influence the database that is being used.

Therefore, it is important to understand the capabilities of the database and tools to support these types of decisions and requirements.

There is a need to understand if something limits the use of the data or creates a need for complex data integrations if the data is going to be leveraged in other places.

It is good to recognize that the data requirements for one application might not be the same for another one, and being a database administrator, there are opportunities here to help develop a data architecture or plan to be able to reuse data, leverage the information in other analysis and calculations where it makes sense, or even use the data in other formats, making it easy to create specific applications that might be different than relational models.

So, if you need to use different models of the data, such as hierarchical, graph, and object, are there ways you can do that in the Oracle Database? Of course! The point of this discussion is to look at different ways of using the same data sets and leveraging a relational model for graph models, hierarchical for machine learning, and all of the relational models for several different uses and applications.

Workloads also play a part of how you model your data. Is this transactional data? Warehouse data? Normally, I find there are different hybrid approaches to this as well.

All of these factors come into play when designing systems and wanting to make the applications perform well. There are parameters and configurations to tune the database based on workloads and types of data being used, but database doesn’t limit the workloads or types.

The database architect and administrator have a responsibility to create a robust system and data model for these applications and tune them along the way.

After the applications are running and you’ve collected all kinds of data and information, shouldn’t this data be harnessed to be used for business decisions and other applications? Absolutely. Even after all the model decisions are made and applications have been running, there are reasons to use data in different ways because of the importance of the information that has been collected.

Oracle 23c has several new features that are focused on easier data management, using different data types and capabilities with all different kinds of workloads. This is to help leverage the data that is already there and gather new information, while ideally simplifying development.