Data Exploration and Uploads

Browse your data

Use the editor's Catalogs panel or the catalog browser to inspect available catalogs, schemas, tables, and columns. Results reflect your connection and permissions. Use qualified object names in saved SQL to make its dependencies clear.

The editor's object browser also groups AI/ML objects, including models, vector indexes, and feature groups, and lists Procedures. An empty list can mean no objects exist in the selected context or that access is missing. After creating an object, refresh the listing and check its fully qualified name.

Upload files

Open File Upload to upload CSV, Parquet, JSON, or JSONL files, subject to the limits shown by Studio. An uploaded file is staged input; uploading alone does not create a queryable table.

1. Review the file's columns and intended types.
2. Choose a destination catalog and schema where you may create or write tables.
3. Create the target table with suitable types, if necessary.
4. Use the uploaded file's staging reference with COPY INTO.
5. Inspect the load result, row count, sample values, and any rejected records before using the table downstream.

Do not use paths from your laptop as if they were files on the hosted engine. For external storage, use a supported stage or authorized storage integration.

Explore before changing data

Start with a bounded read and inspect schema and types. Before modifying or dropping objects, confirm the connection and qualified target. Snapshots and time travel explain historical reads and rollback; retention limits still apply.