Skip to main content

Get Dataset Schema

GET 
/datasets/:dataset_id/schema

Get the dataset's current schema: columns, types, nullability, and row count.

Reads the latest version's Parquet directly with DuckDB's DESCRIBE, so it reflects the dataset's actual current content rather than a stored or inferred schema that could have drifted from it. Also reports the schema's row-identity column when the data still carries it — a prior column mapping can rename it away, in which case this comes back null so a caller does not build a query keyed on a column that no longer exists.

Request​

Responses​

Successful Response