Schema

Schemas are automatically discovered by SWE once the initial bulk of data is sent to the dataset. They can also be created manually via the UI or SDK.

Regardless of the discover/create mechanism used initially schema may be edited by users at any time. However, note that changes, especially to data types, may result in historical data loss.

Once a schema exists, SWE enforces it for any new data ingested into the dataset to ensure that:

• Any undefined field is ignored.
• Field type mismatches trigger errors (registered as Events).
• Omitted fields from the schema are inserted with their default values.