Graphed LogoGraphed
Graphed LogoGraphed
How It WorksAgentsIntegrationsPricing

What is F64 in Power BI?

Cody Schneider••9 min read

If you've spent any time getting your data ready in Power BI's Power Query Editor, you've probably seen a variety of data type icons and names mentioned. One term that often pops up in error messages, diagnostic logs, or technical discussions is F64. While you might not click a button labeled "F64," understanding what it is, where it comes from, and how it impacts your reports is a foundational skill. This article breaks down what F64 really means, why it matters, and how to make sure you’re using it - and its alternatives - correctly.

GraphedGraphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Watch Graphed demo video

So, What Exactly is F64 in Power BI?

Put simply, F64 is the low-level, technical name for a "64-bit double-precision floating-point number." That’s a bit of a mouthful, so let's unpack it.

  • Number: This part is easy. It’s used to store numerical data, not text or dates.
  • Floating-Point: This is a key detail. It means the number can have a decimal point that "floats." This versatility allows it to represent fractional values like 123.45, scientific notation like 1.23e+8, and very small numbers like 0.00000987. This is in contrast to an integer, which can only store whole numbers (e.g., 5, 100, -50).
  • 64-bit: This refers to the amount of computer memory allocated to store the number. A 64-bit number uses more memory than a 32-bit number, but in exchange, it offers significantly more precision (more accuracy in the decimal places) and a much wider range of possible values.
  • Double-Precision: This is another term for "64-bit." In the world of computing, "single-precision" often refers to 32-bit floating-point numbers, while "double-precision" refers to 64-bit.

In the Power BI interface, you won't see "F64." You'll see the friendlier term: Decimal Number. But under the hood, when the Power BI engine or Power Query's M language is working, it's often operating with the F64 or Type.Double data type. Think of it as the engine's internal jargon for what we see as a simple decimal.

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Where Does F64 Actually Show Up?

The primary place you’ll interact with this data type is within the Power Query Editor, even if it goes by another name. When you import data, Power BI often tries to guess the data type for each column in a step called "Changed Type." It's great at identifying numbers with decimals and assigning them the Decimal Number type, which is our trusty F64.

You can see and change this data type easily:

  1. In Power BI Desktop, click Transform data from the Home ribbon to open the Power Query Editor.
  2. Select a query (a table) from the left-hand pane.
  3. Look at a column containing numerical data, like Price, TaxAmount, or ProfitMargin.
  4. At the head of the column, to the left of the column name, you'll see a little icon. For a Decimal Number (F64), this icon is "1.2".

If you click that "1.2" icon, a dropdown menu appears allowing you to change the type. While this menu is user-friendly, the F64 term can also appear in several other places:

  • Error Messages: Sometimes, if you try to perform an operation between mismatched data types (like mixing text and numbers), an error message might say something like, "We couldn't convert the value to type double." Double is a programmer's term for the F64 type.
  • DAX Query View & Diagnostics: More advanced users working in DAX Studio or with Power BI’s Performance Analyzer might see F64 mentioned in the diagnostic files. This tells you how the DAX engine is processing and storing your data in memory.
  • Advanced M Code: If you view the M code behind your Power Query steps in the Advanced Editor, you might see data type declarations like type number or Type.Double, both of which correspond to F64.

Why You Should Care About Data Types Like F64

Choosing the right data type isn't just a technical detail - it has a direct impact on the accuracy of your calculations, the performance of your report, and the size of your Power BI file. Each type has its pros and cons.

The Pros: When F64 is Your Best Friend

  1. Incredible Flexibility: It can handle a massive range of numbers, from enormous values down to incredibly small fractions. This makes it a great general-purpose choice for any numerical column with decimals, such as sensor readings, scientific measurements, or complex ratios.
  2. High Precision: With 64 bits to work with, it can store numbers with many decimal places of precision, which is critical in statistical or scientific analysis.

The Cons: Things to Watch Out For

  1. Performance Impact: F64 numbers consume more memory and CPU resources than simpler types like whole numbers (integers). Inside Power BI, the VertiPaq analysis engine that makes your visuals fast is highly optimized for compressing and working with integers. A column of integers will always perform better than a column of F64s. If you have a column with millions or billions of rows that only contains whole numbers (like ProductID, Year, or UnitsSold) but is incorrectly classified as Decimal Number, changing it to Whole Number can significantly improve your report's speed and reduce your model's size.
  2. The Floating-Point Inaccuracy Problem: This is a classic computer science issue present in almost all programming languages, not just Power BI. Because of how numbers are stored in binary (a system of 0s and 1s), computers cannot represent certain base-10 fractions perfectly. For example, the fraction 0.1 cannot be represented perfectly in binary, much like 1/3 cannot be a clean, terminating decimal (it's 0.33333...). The result is a nearly invisible imprecision. Your value might be stored as 19.9999999999999998 instead of exactly 20. For summing up sales or displaying on a chart, this tiny difference is irrelevant. But if you're writing a DAX measure that checks for exact equality - like IF([Value] > 19.99, "Pass A", "Fail") - a value entered originally as 20 could get evaluated oddly in certain conditions due to the invisible rounding inaccuracies of F64. In those uncommon situations, this can cause major headaches.

F64 vs. The Other Numbers: A Quick Comparison

To really use your data well, you need to choose the right tool for the job. Here’s how F64 (Decimal Number) stacks up against the other two most common numerical types in Power BI.

Decimal Number (Internally F64 / Double)

  • Icon: 1.2
  • Best For: Generic numerical values with decimal points, like percentages, ratios, scientific measurements, or any calculation where you have fractional components. It's Power BI's default 'go-to' for any number with a decimal.
  • Watch Out For: Performance overhead on massive tables and the potential for tiny rounding issues in exact comparisons.

Fixed Decimal Number (Currency)

  • Icon: $
  • Best For: Money! If a column represents SalesAmount, Profit, Cost, or any financial value, this is almost always the correct data type.
  • Here's the Key Difference: Unlike the 'floating' decimal point in F64, this type has a fixed number of four decimal places. Internally, it treats the value as a large integer, which allows it to completely avoid those weird floating-point rounding errors. Summing up two Fixed Decimal Number values will always result in a precise answer, down to the cent.

Whole Number (Internally Int64)

  • Icon: 123
  • Best For: Any number that will never, ever have a decimal. This includes identifier (ID) keys, counts (UnitsSold), categories represented as numbers (ProductCategoryID), and years.
  • Top Benefit: Performance and efficiency. The VertiPaq engine in Power BI loves integers. It can compress them far more effectively, which translates to a smaller file size and faster processing times for your calculations and visuals.
GraphedGraphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Watch Graphed demo video

Best Practices for Managing Your Numbers

Armed with this knowledge, you can now optimize your data model. Follow these simple rules to build faster and more accurate reports.

1. Apply Data Types Early and Deliberately in Power Query

Don't just rely on Power BI's automatic "Changed Type" step. Go through your columns and consciously choose the best type. Is this an ID column mistakenly set as Decimal Number? Change it to Whole Number. Is this column storing price? Change it to Fixed Decimal Number.

2. ALWAYS Use Fixed Decimal Number for Money

This is the most important rule. If your column contains currency, change its data type to Fixed Decimal Number. This will save you from hard-to-diagnose rounding errors when your totals seem off by a fraction of a cent.

3. Convert to Whole Number Aggressively

Be a ruthless optimizer. If a column contains only integers, make sure it’s a Whole Number. Many systems export data as-text or as a generic number type, but you know your data best. Converting these columns makes your model leaner and faster.

4. Use Decimal Number as a Great Default for Everything Else

If your data is not money and not a whole number, then Decimal Number (F64) is the perfect choice. It’s flexible and precise enough for nearly every other business scenario, from calculating engagement rates to tracking physical measurements.

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

5. Be Wary of Direct Equality Checks in DAX on F64 Measures

If you have to write DAX that tests if a Decimal Number column is exactly equal to another decimal number (e.g., IF([Ratio] = 0.25, ...)), consider wrapping it in a ROUND function (IF(ROUND([Ratio], 2) = 0.25, ...)). This neutralizes any potential floating-point inaccuracies by comparing the values at a defined precision level.

Final Thoughts

Ultimately, F64 is just Power BI’s internal name for the versatile Decimal Number data type. While it's a powerful tool for handling fractions and a wide range of values, knowing its limitations is key. By selecting Fixed Decimal for financial data and Whole Number for integers, you'll create a Power BI model that's not only more accurate but also faster and more efficient.

Manually managing data types and optimizing reports across dozens of platforms can turn into a serious time sink. At Graphed, we help you replace that busywork with instant answers. We believe you should spend your time acting on insights, not tidying data columns. Simply connect your various marketing and sales sources, and ask questions like, "Create a sales dashboard from Shopify showing gross sales, and net sales by discounts, monthly." Instead of opening applications or writing DAX code, you just describe the dashboard you want in your native language, and Graphed gets you a current dashboard, complete with interactive charts tracking live data. Give our automations a try and bring your analytics workflow to warp speed with just simple prompts and conversational questions, all in the time it takes using Graphed today!

Related Articles