Test calculated fields with real examples: Write inputs and expected outputs before checking the app; Use a table to show rule behaviour for zero, blank and negative hours; Re-run tests when rates, conditions or rounding rules change
Image: No-Code Business Guide

Data Modelling

Part of No-code calculations and business rules

Testing a calculated field with known examples

Build known-answer examples for a no-code calculated field, then compare results, displays and decisions after a change.

Write down inputs and expected outputs before checking the app's calculated field. Well-chosen examples expose a wrong field reference, condition or missing-value rule. They also give the next owner a repeatable check after a change.

Write the rule in plain language

Take an illustrative service estimate: multiply hours by the hourly rate, then add a fixed fee. The fee still applies when recorded hours are zero. If hours are unknown, the estimate remains pending; unknown hours are not treated as zero. This is an example specification, not a pricing recommendation.

HoursRateFeeExpected resultReason
2.580402402.5 × 80 + 40
0804040Zero is known; the fee still applies
blank8040PendingHours are unknown
−18040Reject for correctionNegative hours are invalid under this example rule

The table states desired behaviour; it does not report a test in a builder. Where a rule flags results only above a threshold, include examples below, equal to and above it. Record the expected flag as well as the number.

Key Test Cases for Service Estimate Calculation

Blank Hours
Pending
Negative Hours
Reject for correction

Compare the actual routes

Enter each case through the route staff use to create or update a record. Inspect the calculated value, its display and any decision that reads it. Save and reopen the record. If imports or automations change inputs, check those routes too.

Check your product's documentation and the route that recalculates the value; recalculation timing depends on the product and configuration.

Keep a useful result record

For each case, record the rule version, inputs, expected result, observed result and any difference. Add a test record ID when one exists.

A difference may point to the formula, input rule, display or expected example. Have the business owner settle which is wrong before changing the formula to make one example pass.

Re-run the examples when an input field, rate source, condition or rounding rule changes. Add a case for a defect found later. Keep the set small enough for someone to use and understand.

More from Data Modelling