Why should you care?

Running totals on bank statements and “rank products per category” need windows — GROUP BY alone cannot see neighbor rows.

See it live — copy this example

Run in MySQL Workbench or the mysql CLI.

SELECT city, customer_id, full_name,
       COUNT(*) OVER (PARTITION BY city) AS customers_in_city
FROM customers;

What happened?

• Each row keeps its identity while COUNT(*) OVER counts peers sharing the same city.
• No GROUP BY — you still see every customer_id.

Practice next

1. Run query on customers with repeated cities.
2. Compare to GROUP BY city — note lost row detail.
3. Add ORDER BY inside OVER in ROW_NUMBER lesson.
4. SUM(total_inr) OVER (PARTITION BY customer_id) on orders JOIN.
5. ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW for running sum.

Remember

OVER defines window frame. PARTITION BY subgroups rows. Keeps detail rows unlike GROUP BY.

DataFlow city cohort size

Analytics adds customers_in_city beside each row for ML feature export.

Outcome: Data science skips extra GROUP BY pass.