orstubs

Orstubs are a type of compact placeholder object used in data modeling and integration to represent a set of possible values for a single attribute when the exact value is not yet determined. They encode multiple alternatives in a single serialized unit to enable deferred resolution and more flexible schema mapping.

The term combines "OR" indicating alternative options and "stub" indicating a minimal, incomplete record. It originated in data engineering communities around 2015 as a means to handle uncertain attributes during ETL and knowledge graph construction.

An orstub comprises: an identifier, a list of option values, a confidence score, a resolutionPolicy, and provenance metadata. It may also include an expiration timestamp. Resolution policy guides how downstream processes pick a single value or propagate the uncertainty.

Orstubs are used in data pipelines where attributes have multiple possible values, such as geographic codes or product categories with overlapping taxonomies. Example: a location field may carry options ["US-CA","US-NV"] with confidence 0.6 for California vs. Nevada. They enable auditing and staged data enrichment.

They add processing overhead and require careful governance to avoid ambiguity. Adoption is strongest in knowledge graphs and data lakes where uncertainty is common. Compatibility with existing schemas depends on standardization efforts.