Polars from Zero

Overview

Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off

One annual plan covers every course and certificate on Coursera. 40% off for a limited time.

Polars is a fast columnar DataFrame engine built on Apache Arrow, and this course teaches you to use it from Rust to do real data-engineering work. You will configure a Cargo project with the lazy and csv feature flags, load wine-ratings.csv into a typed DataFrame, and learn the difference between eager DataFrames for exploration and lazy LazyFrames for production. You will compose select, filter, slice, sort, group_by, agg, and join expressions, then read explain output to see predicate pushdown and projection pushdown rewrite your query before it runs. Module 2 puts the API to work cleaning a real wine-ratings dataset with documented drop, fill, and normalize rules. Module 3 wires everything into wine-pipeline, three Rust CLI binaries that implement a bronze, silver, gold medallion architecture over a shared SQLite database and export a top-10 grape leaderboard as CSV and JSON. By the end you will have a complete, runnable Rust pipeline you can adapt to any tabular dataset.

Syllabus

Polars Foundations

Polars in Rust over the Apache Arrow columnar memory layout, set against pandas as a reference. Cargo setup with the lazy and csv feature flags, the DataFrame and Series types, the col expression, CSV reading with header inference and schema overrides, and the eager versus lazy execution model with predicate and projection pushdown.

Cleaning and Transforming Wine Data

Apply Polars expressions to wine-ratings.csv. Detect and drop nulls with null_count and drop_nulls, normalize text with str.to_lowercase and str.strip_chars, filter by rating bands, sort with sort_by_exprs and SortMultipleOptions, group_by and agg for averages and counts, and join two frames with inner, left, and outer join types.

Building the Medallion Pipeline

Wire the cleaning and aggregation primitives into wine-pipeline, three Rust CLI binaries that share a Cargo workspace and a single SQLite database. Bronze writes raw_wines from CSV with an ingested_at timestamp. Silver applies the cleaning contract and writes clean_wines. Gold filters by min-rating, groups by grape, and exports a top-10 leaderboard as gold_wines.csv and gold_wines.json.