VOLVE DATABASE

About This Dataset

This repository provides the Equinor Volve field data in Parquet format, converted from the original dataset. The Volve field was Norway's first fully disclosed oil field dataset, offering real-world data for data engineering and analysis workflows.

The data has been structured into three normalized tables: daily metrics, monthly aggregations, and well metadata. Perfect for learning SQL, data engineering, or building analytical pipelines.

Quick Start with DuckDB

Query the Parquet files directly without loading into a database. Here's a simple example using DuckDB:

import duckdb

# Connect to DuckDB (in-memory)
con = duckdb.connect()

# Query daily production data
result = con.execute("""
    SELECT
        w.wellbore_name,
        SUM(d.oil_volume) as total_oil,
        SUM(d.gas_volume) as total_gas,
        SUM(d.water_volume) as total_water
    FROM 'daily_production.parquet' d
    JOIN 'wells.parquet' w
        ON d.npd_wellbore_code = w.npd_wellbore_code
    WHERE d.date BETWEEN '2008-01-01' AND '2008-12-31'
    GROUP BY w.wellbore_name
    ORDER BY total_oil DESC
    LIMIT 10
""").fetchall()

for row in result:
    print(row)

No database setup required. DuckDB reads Parquet files directly and efficiently, making it perfect for exploratory analysis and prototyping.

Database Schema

The dataset consists of three interconnected tables tracking well production metrics at different time granularities.