Examples / Data engineering

Fix it before you schedule it

A pipeline whose SQL has a mistake in it. Use Try it on the query step to see the error, fix it, try again, and only then run the whole pipeline. Nothing is scheduled until you say so.

BeginnerAbout 10 minutesSchedule: Only when you run itSample files: none: it makes its own data

Watch it

More data engineering examples

• Morning data guardrail: Check the overnight payments extract every morning and get the numbers by email: row count against the usual, which regions or plans moved, blanks that appeared — even when every rule passes.
• Turn a billing inbox into a table: Works offline: sample messages in exactly the shape the Read-a-mailbox task produces. Filter the payment processor's daily payout reports, count what arrived, and check none are missing. Swap the first step for a real mailbox connection when you are ready.
• Daily data quality gate: Load a product usage extract, profile it, and stop the pipeline when the data is wrong — blanks, duplicates, out-of-range values, and a total that drifts from history.
• Warehouse query to Excel report: Query your warehouse on a schedule, check the result, and drop an Excel file on a shared drive every morning. Uses a connection you set up in Settings.
• See what a pipeline costs: A deliberately heavier calculation so Usage & cost has something to show: machine time, peak memory and rows produced for each step. Run it a few times, then open Settings → Usage & cost.