Validation Nodes

Validation nodes let you define custom conditions inside your model that flag deviations during validation checks.

Overview

Standard model validation checks syntax and data compatibility, but it cannot detect logically incorrect results. Validation nodes fill this gap: they are a special node type that defines conditions your results must satisfy. When a condition fails, the validation reports it.

When to use it

When formulas can produce results that are technically valid but logically wrong (e.g. negative volumes, margins exceeding 100%)
When you want automated checks for business rules that model validation cannot catch

How it works

Validation nodes work like regular nodes but serve a specific purpose: their result is interpreted as a pass/fail signal during validation.

A result of 0 (or all zeros): the check passes (no issues).
A result other than 0: the check fails (the non-zero value indicates the deviation).

You can use any formula logic to express the condition. Common patterns use IF, MIN, MAX and Comparisons & boolean operators

Adding a validation node

Open your model in the Model Editor.
Click Add Node in the action bar.
Select Validation Node from the node type menu.
Name the node and define its formula.

Common patterns

Check that a value is never negative

Node	Formula	Logic
PositiveVolume	`IF('Volume' >= 0, 0, 'Volume')`	Returns 0 when volume is non-negative (passes). Returns the actual volume when negative (fails and shows the deviation).
PositiveVolume	`MIN(0, 'Volume')`	Equivalent shorthand. Returns 0 when volume is positive, returns the negative value otherwise.

Ensure a value stays within a range

Use IF with combined conditions to check a lower and upper bound at the same time.

Node	Formula	Logic
DiscountInRange	`IF('Discount Rate' >= 0 AND 'Discount Rate' <= 0.5, 0, 'Discount Rate')`	Returns 0 when the discount rate is between 0% and 50% (passes). Returns the actual value when outside that range (fails).

Flag values that should never be zero

Use IF to detect zeros where a positive value is expected.

Node	Formula	Logic
NonZeroHeadcount	`IF('Headcount' != 0, 0, 1)`	Returns 0 where headcount is non-zero (passes). Returns 1 where headcount is zero (fails).

Combine multiple conditions

Use AND, OR to check several rules in one validation node.

Node	Formula	Logic
RevenueConsistency	`IF('Revenue' > 0 AND 'Volume' > 0, 0, 1)`	Returns 0 when both revenue and volume are positive (passes). Returns 1 when either is zero or negative (fails).

Running validation node checks

You can evaluate validation nodes in two ways:

From the Model Editor:

Click the Validation Node Checkup button in the header bar. This runs all validation nodes against the current model state.

From a Workspace:

Validation nodes are also evaluated automatically during Workspace Validation, which checks conditions across all active scenarios.

Limitations & edge cases

Validation nodes only check results. They do not change calculations or outputs of other nodes.
The pass/fail signal is based on the numeric result: zero passes, non-zero fails. Design your formula accordingly.
Validation nodes are evaluated per intersection (every combination of dimension values), so a single validation node can flag multiple issues at different intersections.

FAQ

Do validation nodes affect model results?
No. They only check results. They do not influence the calculations of other nodes.

What is the difference between model validation and validation node checks?
Model validation checks syntax and structure. Validation node checks evaluate whether your custom conditions are met at the result level.

Can I use validation nodes without workspace validation?
Yes. You can run the Validation Node Checkup directly from the Model Editor at any time.