Hex Data Notebooks and Apps Cheat Sheet

Updated 2026-05-23

Next Topic: HR and People Analytics Cheat Sheet

🧠Study flashcards on this topic90 cards · spaced repetition→

Hex is a cloud-based collaborative data platform where SQL, Python, and no-code analyses live side-by-side in notebook projects. Cells chain together through a reactive DAG engine, so upstream changes automatically re-execute dependent cells; finished notebooks publish as polished interactive apps with a single click. This cheat sheet covers everything from cell types and SQL chaining through scheduling, secrets, semantic models, embedded analytics, and the Magic AI suite.

What This Cheat Sheet Covers

This topic spans 15 focused tables and 213 indexed concepts, 90 flashcards. Below is a complete table-by-table outline of this topic, spanning foundational concepts through advanced details.

Table 1: Cell TypesTable 2: SQL Cells — Warehouse SQL vs DataFrame SQLTable 3: Input Parameter CellsTable 4: Reactive Execution Model (DAG)Table 5: Magic AI SuiteTable 6: App BuilderTable 7: Publishing and Sharing AppsTable 8: Scheduling and Triggered RunsTable 9: Secrets, Environment Variables, and ComputeTable 10: Data Connections and Semantic LayerTable 11: Version History and Git IntegrationTable 12: Embedded AnalyticsTable 13: Collaboration FeaturesTable 14: Public API and IntegrationsReferences

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: Cell Types

Every Hex project is built from typed cells, each with its own execution engine and output. Choosing the right cell type for the job — rather than forcing everything into Python — keeps notebooks readable and lets Hex optimize execution.

Type	Example	Description
SQL Cell	`SELECT user_id, SUM(revenue) AS total` `FROM orders` `GROUP BY 1`	• Queries a connected data warehouse or an in-memory DataFrame • result exposed as a named Pandas DataFrame automatically
Python Cell	`import pandas as pd` `df = df_orders[df_orders.revenue > 0]`	• Runs arbitrary Python 3 • all cell outputs (DataFrames, variables, figures) are available to downstream cells
Markdown / Text Cell	`## Analysis Summary` `Revenue grew 12 % QoQ.`	• Renders GitHub-flavored Markdown • used for narrative, headings, and callout text between analytical cells
Chart Cell	(drag columns onto X/Y axes in the no-code chart builder)	• No-code drag-and-drop visualization powered by Vega-Lite • auto-suggests chart types based on data shape
Transform Cell	(Pivot, Filter, Sort, Group configured via UI dropdowns)	• Visual Pivot / Filter / Summarize operations on a DataFrame — no code required • output is a new named DataFrame
Input Parameter Cell	(select "Dropdown" type, set values: "Daily", "Weekly", "Monthly")	Creates an interactive widget (text box, slider, date picker, dropdown, etc.) whose value is injected into downstream SQL and Python cells.
Data Cell — Writeback	(configure target table + connection, map columns)	• Writes a DataFrame back to a data warehouse table • useful for annotation workflows and operational data entry inside apps
dbt Metrics Cell	(select metric + dimensions from dbt semantic layer)	• Queries dbt MetricFlow metrics directly without writing SQL • pulls from the connected dbt Cloud semantic layer

Table 2: SQL Cells — Warehouse SQL vs DataFrame SQL

Hex SQL cells can run against two different engines, and understanding the distinction prevents unexpected query behavior and performance surprises. Switching engines changes where the computation runs, how much data returns, and which SQL dialect is valid.

Type	Example	Description
Warehouse SQL	`SELECT * FROM prod.orders WHERE dt = '{{ date_1 }}'`	• Pushes the query to the connected data warehouse • uses the warehouse's SQL dialect • result streams back as a Pandas DataFrame

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

Table 3: Input Parameter Cells

Input cells turn a static notebook into an interactive app — viewers change a widget value and Hex re-runs only the dependent cells. The widget type determines what the user sees and what data type flows into downstream SQL or Python.

Type	Example	Description
Dropdown	(options: "US", "EU", "APAC"; referenced as `{{ region_1 }}`)	• Single-select list • value is a string • can be populated from a SQL cell result dynamically

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

Table 4: Reactive Execution Model (DAG)

Hex's execution engine automatically infers cell dependencies from variable references and runs cells in the most efficient order — no manual wiring required. Understanding how the DAG works helps diagnose unexpected re-execution or cells that seem to skip.

Concept	Example	Description
Dependency Auto-inference	(Python cell reads `df_sql_1` → Hex links it to the SQL cell that produced it)	• Hex scans variable reads across cells and builds dependency edges automatically • no explicit wiring needed

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

Table 1: Cell Types

Type	Example	Description
SQL Cell	`SELECT user_id, SUM(revenue) AS total` `FROM orders` `GROUP BY 1`	• Queries a connected data warehouse or an in-memory DataFrame • result exposed as a named Pandas DataFrame automatically
Python Cell	`import pandas as pd` `df = df_orders[df_orders.revenue > 0]`	• Runs arbitrary Python 3 • all cell outputs (DataFrames, variables, figures) are available to downstream cells
Markdown / Text Cell	`## Analysis Summary` `Revenue grew 12 % QoQ.`	• Renders GitHub-flavored Markdown • used for narrative, headings, and callout text between analytical cells
Chart Cell	(drag columns onto X/Y axes in the no-code chart builder)	• No-code drag-and-drop visualization powered by Vega-Lite • auto-suggests chart types based on data shape
Transform Cell	(Pivot, Filter, Sort, Group configured via UI dropdowns)	• Visual Pivot / Filter / Summarize operations on a DataFrame — no code required • output is a new named DataFrame
Input Parameter Cell	(select "Dropdown" type, set values: "Daily", "Weekly", "Monthly")	Creates an interactive widget (text box, slider, date picker, dropdown, etc.) whose value is injected into downstream SQL and Python cells.
Data Cell — Writeback	(configure target table + connection, map columns)	• Writes a DataFrame back to a data warehouse table • useful for annotation workflows and operational data entry inside apps
dbt Metrics Cell	(select metric + dimensions from dbt semantic layer)	• Queries dbt MetricFlow metrics directly without writing SQL • pulls from the connected dbt Cloud semantic layer

Table 2: SQL Cells — Warehouse SQL vs DataFrame SQL

Type	Example	Description
Warehouse SQL	`SELECT * FROM prod.orders WHERE dt = '{{ date_1 }}'`	• Pushes the query to the connected data warehouse • uses the warehouse's SQL dialect • result streams back as a Pandas DataFrame

Table 3: Input Parameter Cells

Type	Example	Description
Dropdown	(options: "US", "EU", "APAC"; referenced as `{{ region_1 }}`)	• Single-select list • value is a string • can be populated from a SQL cell result dynamically

Table 4: Reactive Execution Model (DAG)

Concept	Example	Description
Dependency Auto-inference	(Python cell reads `df_sql_1` → Hex links it to the SQL cell that produced it)	• Hex scans variable reads across cells and builds dependency edges automatically • no explicit wiring needed