Define what a valid record looks like
Create typed columns for text, numbers, dates, booleans, and JSON. Use required fields, defaults, uniqueness, and validation rules to make expectations explicit.
ART Datalake / Dynamic tables
Give changing business data a dependable structure.
Turn operational information into records your workflows can use. Define the shape of your data, connect related records, and evolve the structure as the business process develops.
Dynamic tables in practice
Built for the way your data works.
Evolve with a record of change
Column changes create schema versions. Existing records are validated when a typed schema is applied, so incompatible data must be reconciled before the change takes effect.
Connect records to the next action
Model relationships between collections and make record changes available to event-driven workflows. Keep connected data consistent as applications read and update it.
Illustrative workflow
A supplier record that can move the process forward
Keep supplier names, review states, and operational details in a table with defined fields. Link related records and use a change in review state as an input to the next workflow step.
- Define fields and validation
- Create and relate records
- Respond to record changes
Understand the lifecycle.
Table deletion is a recoverable archive; permanent removal is a separate operation. Tables default to indefinite retention. Retention settings must also respect relationship constraints.
A few useful details.
Can the schema change after records exist?
Yes. Schema changes are versioned and validated against the existing data. Incompatible records must be reconciled before a typed schema can be applied.
Are tables the same as object storage?
No. Tables hold structured records and relationships. Object storage holds files organized in buckets and folders. They support different parts of the same workflow.
Start with your workflow
What should your data make possible?
Bring the records, files, and context behind your process. We’ll explore how Datalake fits.