Query any data as it was at any point in time
Every record tracks both when a fact was true in the real world and when the platform learned about it. Corrections, late-arriving data, and restated values are all preserved — nothing is overwritten, nothing is lost.
Track what was known and when, across every dataset.
Native bi-temporal support for "as of" and "as at" queries. Point-in-time position lookups, correction tracking, and regulatory lookback — built into the data layer.
What bi-temporal queries look like in practice
Capabilities
Every record, two time axes, zero configuration
Automatic bi-temporal versioning
Every record is versioned on both business time and knowledge time. No opt-in, no configuration — bi-temporal is the default for every dataset.
Point-in-time SQL syntax
Use AS OF and AS AT query modifiers to specify business date and knowledge date independently. Standard SQL extended with temporal semantics.
Correction chain tracking
Every correction is linked to the original record it amends. Follow the chain from original to latest restatement — with timestamps, users, and reason codes.
Regulatory lookback support
Reconstruct past-state reports exactly as they were filed. Specify the knowledge-time cutoff to see only the data available at the reporting date.
Zero-configuration setup
Bi-temporal is the default, not an opt-in feature. Every record, every dataset, every pipeline — automatically versioned along both time axes.
API and warehouse access
Run temporal queries via SQL in the warehouse, through the REST API, or via Snowflake Data Share. Every access channel supports AS OF / AS AT semantics.
Try a point-in-time query
Run yesterday's report with last week's data — bi-temporal by default, no configuration needed.
