Metadata
- Author: Humans of the Data Sphere
- Full Title:: Humans of the Data Sphere Issue #5 December 10th 2024
- Category:: 🗞️Articles
- Finished date:: 2024-12-10
Highlights
(Anchor modeling • It has some similarities with Data Vault, as it focuses on separating core entities (anchors) from their attributes as well as relationships between entities (known as ties). However, anchor modeling is more highly normalized and is less focused on auditability than Data Vault, with a stronger focus on adaptability and schema evolution. One of the primary goals of Anchor Modeling is to support schema flexibility and evolution without disrupting existing structures. Therefore each attribute gets its own table, new attributes can be added without modifying existing tables, ensuring non-disruptive schema updates. This is a more extreme level of normalization than other methodologies. New attributes or relationships can be added without altering existing tables. Schema changes are handled as non-disruptive extensions. Each attribute and tie table track historical changes using validity period columns (ValidFrom, ValidTo), enabling point-in-time analysis. • Pros: Very adaptable to change. Good support for historical analysis and point-in-time queries. • Cons: Similar to Data Vault, except there is less support in tooling. Due to the one-table-per-attribute, there can be more joins.
(View Highlight)