Databricks Notebooks Cheat Sheet

Updated 2026-04-20

Next Topic: Databricks Optimization Cheat Sheet

🧠Study flashcards on this topic116 cards · spaced repetition→

Databricks notebooks are interactive, multi-language development environments within the Databricks Data Intelligence Platform, designed for collaborative data engineering, analytics, and machine learning workflows. They combine executable code cells with visualizations, markdown documentation, and real-time collaboration features. Unlike traditional Jupyter notebooks, Databricks notebooks provide native integration with Apache Spark, automatic versioning, built-in Git support, serverless compute, and AI-powered coding assistance through Genie Code—making them production-ready tools for both exploratory analysis and automated data pipelines. Understanding magic commands, dbutils, widgets, and the notebook execution model is essential for maximizing productivity.

What This Cheat Sheet Covers

This topic spans 14 focused tables and 153 indexed concepts, 116 flashcards. Below is a complete table-by-table outline of this topic, spanning foundational concepts through advanced details.

Table 1: Notebook Basics and Cell TypesTable 2: Magic CommandsTable 3: Code Development FeaturesTable 4: Databricks Utilities (dbutils)Table 5: Compute and ExecutionTable 6: Data VisualizationTable 7: Notebook Widgets and ParametersTable 8: Collaboration and SharingTable 9: Version Control and Git IntegrationTable 10: File Formats and Import/ExportTable 11: Notebook Workflows and OrchestrationTable 12: Debugging and MonitoringTable 13: Advanced FeaturesTable 14: Performance and Best Practices

Quick IndexSubscribe to unlock

A jump-to index of every table row in this cheat sheet.

Mind MapSubscribe to unlock

An interactive map of every table and concept in this topic.

Table 1: Notebook Basics and Cell Types

Start here to understand how a notebook actually runs: code versus markdown cells, the keyboard shortcuts that execute them, and the display() function that gives Databricks its richer table and chart rendering. A couple of less-obvious touches live here too — the notebook's default language, per-cell language overrides, and the _sqldf variable that quietly hands a SQL cell's results to the next Python cell.

Concept	Example	Description
Code cell	`df = spark.read.table("sales")` `display(df)`	• Default cell type that executes code in the notebook's default language (Python, Scala, SQL, or R) • results appear below the cell immediately after execution.
Run cell	Press `Shift+Enter`	• Executes current cell and automatically moves focus to the next cell • `Ctrl+Enter` runs cell without advancing • cells execute sequentially by default.
display() function	`display(spark.read.parquet("/data"))` `display(dbutils.fs.ls("/mnt"))`	• Databricks-specific function for enhanced table rendering with built-in chart builder and data profile • triggers visualization options for DataFrames • shows formatted tables for dbutils commands.
Run all	Click Run All button	• Executes all cells from top to bottom • on DBR 14.0+ runs as a batch workflow and halts on first error • cells with `%skip` are excluded.
Cell output	`display(df)` returns interactive table	• Rendered results below cell • supports tables, charts, HTML, images, and text • output persists in IPYNB format when output commits are enabled.
Markdown cell	`%md` `# Sales Report` `Data updated: 2026-04-20`	• Creates formatted documentation using Markdown syntax • supports headings, lists, tables, images, and inline code for narrative explanations.
Default language	Created as Python notebook	• Language assigned when notebook is created • determines which language cells execute without magic commands • changeable via File → Change default cell language.
Language override	`%sql` `SELECT * FROM sales WHERE region = 'West'`	• Executes cell in specified language regardless of notebook default • supports `%python`, `%scala`, `%sql`, and `%r` magic commands at cell start.
Run selected text	Highlight code → `Ctrl+Shift+Enter`	• Executes only the highlighted lines within a cell without running the full cell • if no text is highlighted, runs the current line • useful for iterating on code quickly.
Multiple outputs per cell	`display(chart)` `print("summary")` `display(table_df)`	• Python cells support multiple display() calls and print statements in a single cell • all outputs render sequentially below the cell.
Paste images	`Ctrl+V` in a Markdown cell	• Paste images directly from clipboard into Markdown cells • also supports drag-and-drop of image files • images stored as notebook attachments.
_sqldf auto-variable	`%sql SELECT * FROM sales` Next cell: `_sqldf.filter("amount > 100")`	• SQL cell results automatically available as a DataFrame named _sqldf in subsequent Python and SQL cells • enables cross-language data sharing without temp views.

Table 2: Magic Commands

The %-prefixed commands that change what a cell does — switch its language (%sql, %python, %scala, %r), render markdown, install packages with %pip, drop into a shell with %sh, or pull in another notebook with %run. The bottom rows surface lesser-known IPython magics like %%writefile and %%capture that quietly expand what a single cell can accomplish.

Command	Example	Description
%sql	`%sql` `SELECT user_id, COUNT(*) FROM events GROUP BY user_id`	• Executes cell as SQL query • results automatically rendered as interactive table with chart options • widget values accessible using `${variable}` syntax.

The rest of this section is part of the full cheat sheet

Table 3: Code Development Features

The editor tooling that makes writing code in a notebook feel less like a stripped-down text box and more like an IDE. Genie Code — the AI assistant renamed in March 2026 — anchors this with autocomplete, agent mode, and slash commands like /fix and /explain, sitting alongside familiar conveniences such as Black formatting, multi-cursor editing, code folding, and a VS Code–style command palette.

Feature	Example	Description
Genie Code	Click Genie Code icon or press `/` in cell	• AI-powered code generation, debugging, and explanation (renamed from Databricks Assistant, March 2026) • understands notebook context and Unity Catalog metadata • generates code from natural language descriptions.

The rest of this section is part of the full cheat sheet

Table 4: Databricks Utilities (dbutils)

dbutils is the built-in toolbox you reach for constantly — file operations on dbutils.fs, secret retrieval that auto-redacts in output, notebook-to-notebook orchestration with run() and exit(), and the restartPython() you'll need after a %pip install. Note mount() is marked legacy, since Unity Catalog volumes are now the recommended path.

Utility	Example	Description
dbutils.fs.ls()	`dbutils.fs.ls("/mnt/raw-data/2026/")` `display(dbutils.fs.ls("dbfs:/tmp"))`	• Lists files and directories at specified path • returns FileInfo objects with name, path, size, and modification time • works with DBFS, mounted storage, and Unity Catalog volumes.

The rest of this section is part of the full cheat sheet

Table 1: Notebook Basics and Cell Types

Concept	Example	Description
Code cell	`df = spark.read.table("sales")` `display(df)`	• Default cell type that executes code in the notebook's default language (Python, Scala, SQL, or R) • results appear below the cell immediately after execution.
Run cell	Press `Shift+Enter`	• Executes current cell and automatically moves focus to the next cell • `Ctrl+Enter` runs cell without advancing • cells execute sequentially by default.
display() function	`display(spark.read.parquet("/data"))` `display(dbutils.fs.ls("/mnt"))`	• Databricks-specific function for enhanced table rendering with built-in chart builder and data profile • triggers visualization options for DataFrames • shows formatted tables for dbutils commands.
Run all	Click Run All button	• Executes all cells from top to bottom • on DBR 14.0+ runs as a batch workflow and halts on first error • cells with `%skip` are excluded.
Cell output	`display(df)` returns interactive table	• Rendered results below cell • supports tables, charts, HTML, images, and text • output persists in IPYNB format when output commits are enabled.
Markdown cell	`%md` `# Sales Report` `Data updated: 2026-04-20`	• Creates formatted documentation using Markdown syntax • supports headings, lists, tables, images, and inline code for narrative explanations.
Default language	Created as Python notebook	• Language assigned when notebook is created • determines which language cells execute without magic commands • changeable via File → Change default cell language.
Language override	`%sql` `SELECT * FROM sales WHERE region = 'West'`	• Executes cell in specified language regardless of notebook default • supports `%python`, `%scala`, `%sql`, and `%r` magic commands at cell start.
Run selected text	Highlight code → `Ctrl+Shift+Enter`	• Executes only the highlighted lines within a cell without running the full cell • if no text is highlighted, runs the current line • useful for iterating on code quickly.
Multiple outputs per cell	`display(chart)` `print("summary")` `display(table_df)`	• Python cells support multiple display() calls and print statements in a single cell • all outputs render sequentially below the cell.
Paste images	`Ctrl+V` in a Markdown cell	• Paste images directly from clipboard into Markdown cells • also supports drag-and-drop of image files • images stored as notebook attachments.
_sqldf auto-variable	`%sql SELECT * FROM sales` Next cell: `_sqldf.filter("amount > 100")`	• SQL cell results automatically available as a DataFrame named _sqldf in subsequent Python and SQL cells • enables cross-language data sharing without temp views.

Table 2: Magic Commands

Command	Example	Description
%sql	`%sql` `SELECT user_id, COUNT(*) FROM events GROUP BY user_id`	• Executes cell as SQL query • results automatically rendered as interactive table with chart options • widget values accessible using `${variable}` syntax.

Table 3: Code Development Features

Feature	Example	Description
Genie Code	Click Genie Code icon or press `/` in cell	• AI-powered code generation, debugging, and explanation (renamed from Databricks Assistant, March 2026) • understands notebook context and Unity Catalog metadata • generates code from natural language descriptions.

Table 4: Databricks Utilities (dbutils)

Utility	Example	Description
dbutils.fs.ls()	`dbutils.fs.ls("/mnt/raw-data/2026/")` `display(dbutils.fs.ls("dbfs:/tmp"))`	• Lists files and directories at specified path • returns FileInfo objects with name, path, size, and modification time • works with DBFS, mounted storage, and Unity Catalog volumes.