Skip to content

Automatic translation — the Portuguese version is the reference. View in Portuguese

Updated on 2026-10-03

Data console screen — write and run SQL with the cost upfront

The SQL console in the Data menu (/dados) — table explorer, Run, Estimate, View schema, Reveal PII, export, and history — with the literal label of each button. Plus the index of the other data screens.

On this page (9)

This page is an interface inventory: the text here is the text on screen, literally.

Where it is: menu Data → Console (address /dados). The screen title is Data, and the subtitle says the main thing: "Write, run and see the result of your SQL queries — with a cost per run. Every query is billed."

It's the screen for people who know (or are learning) SQL and want to look at the datalake tables directly. If you'd rather write the question in plain language, use the Queries screen.

The other data screens#

Until 10/03/2026, this page covered seven screens on its own. Now each screen has its own:

Run your first query#

  1. 1. Find the table in the Explorer

    On the left is the Explorer, with the datasets and tables. Use the "Search table…" search to filter. Click a table to select it.

  2. 2. Check the columns

    With the table selected, View schema opens the columns and types (the panel also shows Rows, Size, Created, and Updated). Query table writes a ready-made SELECT in the editor.

  3. 3. See the cost first

    Click Estimate: the screen shows how much the query will read and how much that costs — without running and without charging.

  4. 4. Run it

    Click Run (shortcut Ctrl+E; while running: "Running…"). The result appears below, in the Result tab.

  5. 5. Save it, if you'll use it again

    Give it a name in Query name and click Save. The query goes to the Saved tab.

Data console with a query run: the Explorer on the left, the editor with Run and Estimate, and the Result tab with the run's cost

What each button does#

You want toClick
Run the queryRun (shortcut Ctrl+E)
Know the cost firstEstimate
Save the querySave (with the Query name next to it)
See a table's columns and typesView schema → Query table
Open another query without losing the current oneNew query (opens a new tab)
Compare two queries side by sideSplit (becomes Merge)
Create a datasetCreate dataset → Create
See personal data in clear textReveal PII (becomes Hide PII)
Take the result elsewhereExport → Confirm and download CSV

The bottom panel has three tabs: Result · History · Saved. In the Saved tab, Load in editor brings the query back. The result comes in pages (Previous · Next, and Rows per page).

Data console with a table's schema open: columns, types, rows, size, and the Query table button

Every run is billed

The amount depends on the volume of data the query reads, not on the number of rows returned. That's why Estimate exists: it's free. When the same query was run a short while ago, the result may come from the cache, marked as "⚡ cache · free".

When the SQL changes data#

The product asks for confirmation when the SQL changes or deletes data

Before running, the screen asks in no uncertain terms: "⚠ This CHANGES data — confirm?" or "⚠ This DELETES data — confirm?", and the confirmation button repeats the action: Confirm and CHANGE / Confirm and DELETE. There's no path where a write happens without this step — and it can't be undone.

Personal data (PII)#

By default, columns marked as personal data come masked: "PII data is anonymized in the query — only in our layer, without changing the datalake." Those with permission can click Reveal PII; the screen then warns "PII data REVEALED in this view — access logged in the audit trail." To go back, Hide PII.

Who can reveal is decided on the Data governance screen.

Export the result#

Export opens the Export data warning: "The customer is responsible for exported data. Make sure it is handled properly under the Brazilian General Data Protection Law (LGPD) and your internal policies." The file comes out as CSV when you click Confirm and download CSV.

A large result is a sample

If the query produces more rows than the grid loads, the screen marks the result as Partial result: the totals on screen aren't the whole result. Aggregate in the SQL itself (SUM, GROUP BY) to see the complete number.

When the screen is empty#

  • No tables: "No datasets yet. Create pipelines and transformations to populate the datalake." — the first step is a pipeline.
  • No run: "Run a query to see the result."
  • Empty history: "No runs yet. Run a query to start the history."
  • Nothing saved: "No saved queries. Use “Save” to reuse queries."

In a demo environment, the demo badge appears: nothing is actually read from the datalake.

Back to the app#

Next task#

Related links