# Parquet-ly Cloudy

THE CLOUD EXPEDITION! · level 13 · chapter 5

Aria's eyes go wide behind her spectacles. 'Forget CSV — Parquet is the FORMAT OF LEGENDS. Columnar storage, built-in compression, blazing-fast scans.' She fans herself with a schema diagram. 'DuckDB devours Parquet files straight from cloud storage — point it at a URL and it reads only the columns you ask for.' Time to discover <span class='feature-highlight'>Parquet Files</span> — reading columnar data directly from cloud storage.

## Task

Use `read_parquet()` to read the 2025 regional census archive and summarize total population by region. Return <b>region</b> and <b>total_population</b>, most populated regions first. File: `https://data.dbquacks.com/archives/census_2025.parquet` (public Parquet). Columns include `region` and `population`.<br><br><div class="text-xs text-gray-400 mt-2">Regional tip: MotherDuck runs in US East, US West, and EU Central. For best performance, query data stored in your region. This file is hosted in US East.</div>

## Hint

read_parquet() accepts an https:// URL directly — no need to download or attach the file

## Expected result

Rows with columns: region and total_population, one per region, ordered by total_population descending

## Tables

### `lake_visits` (500 rows)

| column | type |
| --- | --- |
| visit_id | BIGINT |
| duck_id | BIGINT |
| destination_pond | VARCHAR |
| visit_date | DATE |
| purpose | VARCHAR |

### `pondville_events` (50 rows)

| column | type |
| --- | --- |
| event_id | BIGINT |
| event_name | VARCHAR |
| event_date | DATE |
| organizer_id | BIGINT |
| location | VARCHAR |
| attendee_count | BIGINT |

### `pondville_residents` (100 rows)

| column | type |
| --- | --- |
| duck_id | BIGINT |
| name | VARCHAR |
| species | VARCHAR |
| home_pond | VARCHAR |
| age | BIGINT |
| occupation | VARCHAR |
| joined_date | DATE |

## Data

The tables live in a DuckDB database: https://dbquacks.com/data/databases/motherduck_tutorial.duckdb

To query it outside the browser, attach it read-only:

```sql
ATTACH 'https://dbquacks.com/data/databases/motherduck_tutorial.duckdb' AS db (READ_ONLY);
USE db;
```

This level also queries MotherDuck shares, which need a MotherDuck token. On the page, the editor asks for one.

Play it in the browser: https://dbquacks.com/motherduck/13

## Reference

https://duckdb.org/docs/stable/data/parquet/overview