liveBook · Manning

Columnar storage engine

As explained in Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI

A columnar storage format stores values from the same table column contiguously to support efficient column-oriented access.

A columnar storage format keeps values from the same table column together, making dense tabular data more efficient to store and access than with row-based storage. section 2.3.1

Polars uses columnar storage alongside high-performance execution, improved memory management, and large-dataset handling. Its DataFrame API resembles pandas, while eager and lazy modes support different execution patterns. section 2.3.1

Choosing a tabular-data engine

For massive or dense tabular datasets, evaluate Polars alongside pandas and Spark. Check whether required pandas functions are available before refactoring, then compare the options against dataset size and usage. section 2.3.1

How Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI covers it

How deep it goes
Defines columnar storage and contrasts it with row-based storage, but does not descend into storage-engine internals.
Examples
Uses Polars, pandas, Spark, dense tabular data, and large-dataset tool selection as practical examples.
What it assumes
Assumes a tabular-data context and pandas usage while defining columnar and row-based storage in practical terms.
What you can do after
Evaluate whether Polars, pandas, or Spark better fits a dataset’s size, access pattern, and existing API requirements.
How it connects
Connects columnar storage with Polars, pandas, DataFrame libraries, and row-based storage.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI. The section links point to the parts of the book it draws on.