#PowerPlatformTip 155: 'Float Before Div for Correct Decimals in Power Automate'

TL;DR: In Power Automate div() truncates when both operands are integers. Cast them with float(), div(float(a), float(b)), so the result is a Float and your decimals survive.

You build a percentage or ratio from two whole numbers and the result is 0 or a rounded-down integer instead of a proper decimal. The reason is simple: div() returns an Integer whenever both operands are integers, it truncates, it does not round. Convert the operands to Float first and everything lines up.

💡 Challenge

You want a completion or share percentage, for example matching records versus all records. A natural expression looks like this:

mul(div(outputs('Count_Matching'), outputs('Count_All')), 100)

With Count_Matching = 75 and Count_All = 200 you expect 37.5, but you get 0. Both outputs are integers, so div(75, 200) truncates to 0 and mul(0, 100) stays 0. Whenever the dividend is smaller than the divisor the result collapses to zero, and larger values silently lose their decimals.

✅ Solution

div() only returns a Float when the dividend or the divisor is a Float. Wrap both operands in float() so the division keeps its decimal places:

mul(div(float(outputs('Count_Matching')), float(outputs('Count_All'))), 100)

With the same inputs this now returns 37.5. One Float operand is technically enough, but casting both is explicit and safe when the source type can vary.

🔧 How it’s done

1. Spot the integer division.

🔸 Look for div() where both operands are whole numbers, for example Compose outputs, length() results, counts or integer variables.

🔸 Quick test in the expression editor: div(11, 5) returns 2 (not 2.2) and div(75, 200) returns 0.

2. Cast both operands to Float.

🔸 Change div(a, b) to div(float(a), float(b)):

mul(div(float(outputs('Count_Matching')), float(outputs('Count_All'))), 100)

🔸 As soon as one operand is a Float, div() returns a Float, e.g. div(float(75), float(200)) returns 0.375.

Real flow run: @div(7, 2) with two plain integer operands truncates to 3, the .5 is simply gone.

Real flow run, same numbers: @div(float(7), float(2)) correctly returns 3.5, casting both operands to Float is the entire fix.

3. Format the output for display (optional).

🔸 Wrap the result in formatNumber(..., 'N2') for a clean two-decimal value:

formatNumber(mul(div(float(outputs('Count_Matching')), float(outputs('Count_All'))), 100), 'N2')

🔸 This returns 37.50 as a string. If you keep calculating on it, parse it back with float(formatNumber(...)).

4. Guard against divide-by-zero.

🔸 div() fails when the divisor is 0. Protect the expression with if():

if(equals(outputs('Count_All'), 0), 0, mul(div(float(outputs('Count_Matching')), float(outputs('Count_All'))), 100))

🔸 A zero total now returns 0 instead of throwing an error.

🎉 Result

Your percentage and ratio calculations return accurate decimals instead of 0 or truncated integers, no Azure Function and no extra actions, just the built-in float() conversion, optionally formatted with formatNumber().

🌟 Key Advantages

🔸 Correct math: decimals survive instead of being truncated to Integer.

🔸 No workaround actions: a pure expression fix, no extra Compose or Scope needed.

🔸 Locale-safe display: formatNumber(..., 'N2') gives predictable, formatted output.

🔸 Robust: combined with an if() guard it also handles divide-by-zero cleanly.

🛠️ FAQ

Q1: Why does div() drop my decimals?

If both operands are integers, div() returns an Integer and truncates. It only returns a Float when the dividend or the divisor is a Float, so div(11, 5) is 2 but div(11.0, 5) is 2.2.

Q2: Do I have to cast both values, or is one enough?

One Float operand is technically enough to force a Float result. Casting both with float() is more explicit and protects you when the input type can change.

Q3: Why is my result a string after formatting?

formatNumber() returns a string. If you need to keep doing math, parse it back with float(formatNumber(...)).

Q4: How do I avoid a flow failure when the divisor is 0?

Wrap the expression in if(equals(divisor, 0), 0, div(...)) so a zero total returns 0 instead of throwing an error.

🔗 Related Tips

• #PowerPlatformTip 59: Handle Dynamic Content, safely access properties in expressions with the ? operator.
• #PowerPlatformTip 3: Advanced Filtering Array, write custom expressions in Filter Array’s Advanced Mode.
• #PowerPlatformTip 145: Power Platform Tools, browser-based helpers for building and translating Power Platform expressions.