Skip to main content

Query your data

Where: sidebar → Query.

Access and limits​

The workbench executes read-only SQL against the platform query service. It is not an SQL console for modifying source systems. Use supported pipeline and catalog operations for data changes.

In the reviewed backend, submitting or cancelling a query is a workspace-write operation. Organization owners, administrators, and data engineers have that role permission; analysts and viewers do not. Catalog permissions are an additional requirement, not a replacement for the role check.

Set the query context

Select an available catalog and, when needed, a schema. Use names from the Data Catalog, not guessed names from a source connection.

Write a bounded query

Begin with a small SELECT and a LIMIT. For example, replace the illustrative table name below with a table in your workspace.

SELECT order_id, status
FROM iceberg.training.orders
LIMIT 20;

If you have no dataset yet, use SELECT 1 AS example.

Run and inspect

Select Run query. Wait for completion, then review the result grid, statistics, and any failure details. The server caps returned rows; a displayed sample is not a full-table export.

Revisit the execution

Select an entry in History to inspect its SQL, status, and available results. Use Cancel on an active execution when offered, then verify its final status.

Troubleshooting​

  • Run query is disabled: wait for catalogs to load, select a catalog, and enter SQL.
  • Permission denied: ask your administrator to check both your role and named-resource grants.
  • Table or column not found: inspect the current catalog structure and query context.
  • Updates stopped: reopen the history entry before resubmitting. A lost browser connection does not establish that the query stopped.
  • Mutation rejected: this workbench is read-only. Do not try to evade validation by wrapping a write statement in another expression.