Function Nodes

Function nodes are the workhorses of a Flow. They sit between your data sources and outputs, shaping data as it passes through. You add them from the Functions section of the Element Panel.

Pick a node by what it changes

The easiest way to choose a Function node is to think about what part of the data it changes — rows, columns, or structure:

The node...	...changes	Node
Removes rows	rows	Filter
Adds or changes columns	columns	Formula
Rolls rows up into totals	rows	Group By
Reshapes rows into a new (nested) structure	structure	Map

The other Function nodes are variations on these themes — ways of selecting, combining, or expanding rows.

All Function nodes

I want to...	Use
Keep only rows that match a condition	Filter
Reorder rows	Sort
Remove duplicate rows	Distinct
Add or change column values	Formula
Add columns from another source by matching keys	Lookup
Join two datasets on a key	Merge
Stack rows from multiple sources	Append
Calculate totals, averages, counts per group	Group By
Expand array values into individual rows	Flatten
Reshape rows into a different (nested) structure	Map
Classify text with AI	Classify (see below)

The expression language

Filter, Formula, Group By, and Map all evaluate small expressions to compute values or test conditions. They share one strict-SQL style Expression Language — learn it once and it works across all of them.

A note on Map, Formula, and Group By

Map is the most general data-shaping node: it can place, rename, and compute fields, build nested arrays, and summarize child groups. Formula and Group By are focused, everyday versions of that same capability — Formula computes columns, Group By summarizes rows — and all three author their expressions with the same fx language. Reach for Formula or Group By for the common cases, and Map when you need to change the shape of the data itself.

The Classify node

The Classify node uses an AI connection to categorize text — for example, tagging free-text notes as Mechanical, Electrical, or Process. It takes two inputs (the records to classify and the category definitions) and assigns each record to the closest-matching category by AI embedding. Unlike the editor-time AI assistants (Copilot and Formula Assist), Classify runs as part of flow execution. See the Classify Node for full configuration.

Next steps

• Expression Language — operator and function reference
• Formula — add or change columns
• Filter — keep rows that match a condition
• Map — reshape rows into a new structure
• Building Flows — put nodes together into a Flow