Median delivery time per city

Constraints

1. 1This engine has no MEDIAN and no PERCENTILE_CONT. Build it from a window function.
2. 2Two cities have an odd number of deliveries and one has an even number. Your query must handle both without a special case: with an odd count the median is the single middle row, with an even count it is the average of the two middle rows.
3. 3deliveries must report how many deliveries the city actually made — not how many rows survived your filter.
4. 4Round median_minutes to one decimal place, so the even-count city reads honestly.
5. 5Order by city.

Worked example

Chennai has eight deliveries: 19, 23, 27, 30, 38, 52, 71, 96. There is no single middle row, so the median is the average of the 4th and 5th — (30 + 38) / 2 = 34.0. Bengaluru has nine: 18, 22, 26, 28, 30, 30, 34, 41, 150. The 5th row is the middle, so the median is 30.0 — and its mean is 42.1, because that 150 has to go somewhere. The trick that covers both: take the rows at positions (n + 1) / 2 and (n + 2) / 2 using integer division. When n is odd those two expressions name the same row, so averaging is a no-op. When n is even they name the two straddling rows.

What this tests

Whether you can express a median without a median function — a standing interview question, because at least one engine every team uses is missing one. Also whether you noticed the even-count case at all.