Optimizing Numeric Data Types with Oracle to Google Cloud Migration
Recently, I had the opportunity to present a lightning talk at PGConf India 2023 on the topic of how to prevent the Number data type from becoming a villain during your migration journey.
The talk I gave at PGConf India 2023 was focused on providing a guide for organizations migrating from the legacy commercial engine (Oracle) to Google Cloud, with a specific emphasis on the NUMBER data type. The key takeaways were as follows:
- Providing guidance on best practices for data types during the migration process of relational databases
- Enabling customers to refactor schema during the conversion phase by selecting the optimal numeric family and avoiding failure in later stages
- Demonstrating an automated method for obtaining recommended data types from Oracle when the number declaration lacks precision or scale.
Oracle Current state.
Depending on the use case, Oracle as a source database can be migrated to various data platforms on Google Cloud. It’s crucial to consider optimal data types for each target platform, as it can have a significant impact on performance and storage.
The declarations for the NUMBER data type in Oracle can vary and be dynamic in nature. They may be defined without precision or scale, with precision only, with both precision and scale, or even with negative scale.
Overall it becomes critical to choose optimal data type on any prefer target on Google cloud.
Automated Approach
Check out the GitHub repository for guidance on obtaining optimal recommendations for the NUMBER data type on your preferred target platform on Google Cloud.
Presentation Slide
This presentation deck provides a walkthrough on how to address the Number data type recommendation for a PostgreSQL database on Google Cloud using either Cloud SQL or AlloyDB.
