The actual process of converting decimals to fractions
I keep seeing people struggle with this in spreadsheets and code reviews. The method is straightforward if you understand what's actually happening under the hood. Take 0.75 for example. You write it as 75 over 100, then reduce by finding the greatest common divisor. Both numbers divide evenly by 25, giving you 3/4. That's it. Nothing fancy. The part most people skip is understanding when the straightforward approach breaks down. Recurring decimals exist. Things like 0.3333... or 0.1666... don't behave the same way. I spent about three weeks last year debugging a financial calculation where someone's code was treating 0.3333333333 as exactly 1/3. It wasn't. The floating point representation introduces tiny errors that compound when you're working with currency values across thousands of transactions. I ended up writing a custom routine that detects repeating patterns first, then applies the algebraic method for recurring decimals instead of the standard GCD reduction.
Turn Decimal Into Fraction using the GCD method
For terminating decimals, here's the practical approach. Count the decimal places. Move the decimal point right by that many positions to get your numerator. Your denominator is just 10 raised to the power of however many places you moved. Then find the GCD and divide both sides. A number like 2.125 has three decimal places, so you get 2125 over 1000. The GCD of those two numbers is 125. Divide everything out and you're at 17/8. The edge case nobody warns you about involves decimals that look terminating but aren't. I ran into this when a colleague passed me a value from a database query that looked like 0.5714285714285714. My gut said 4/7, but the raw number isn't exactly 4/7. It's a double-precision float approximation. If you blindly apply the GCD method to this, you get something ugly like 4010293/7016817. The workaround is to set a tolerance threshold. Compare the decimal against known fractions within a small epsilon value, like 0.0001, and snap to the closest match. It saves you from generating absurdly complex fractions that no one can work with. Another thing that catches people out is mixed numbers. The decimal 3.6 converts to 36/10 which reduces to 18/5, but that's an improper fraction. In practice you usually want 3 and 3/5. You pull out the whole number part before reducing, or you convert after. Both work. Just be consistent.
There are also cases where the straightforward method produces fractions so large they're useless. Take pi divided by four, approximately 0.785398163397. The exact fractional representation through continued fractions gives you increasingly accurate approximations. 3/4 is decent. 19/24 is better. 3927/5000 is what you'd get from the naive method, and it's nearly exact but completely unwieldy. Continued fraction expansions let you truncate at whichever precision level makes sense for your use case. If you need to do this conversion frequently, there are a few programs worth checking out. WolframAlpha handles virtually any decimal you throw at it and shows the steps. For spreadsheet work, Excel doesn't have a built-in function, but you can construct one using GCD. Google Sheets has a slightly better experience with its FRACTION function. Python users have the fractions module built in, which handles simplification automatically and deals with recurring decimals through the limit_denominator parameter.
Get the Full Details
When the method hits its limits
Not every decimal converts cleanly. Irrational numbers like sqrt(2) or e don't have exact fractional representations. You can approximate them, but the approximation is always just that. The further you push the precision, the larger the numbers get. There's a tradeoff between accuracy and readability that you have to decide on case by case. Some decimals also come from measurements with inherent uncertainty. If someone tells you a length is 2.3 meters, treating that as exactly 23/10 implies more precision than actually exists. The decimal might represent a range. Converting it to a fraction doesn't magically make it more exact. This matters in engineering and science contexts where significant figures carry meaning. The biggest practical limitation I've encountered is in software that stores decimals as binary floating point. Values like 0.1 and 0.2 don't have exact binary representations. They're repeating fractions in base 2, just like 1/3 is a repeating decimal in base 10. So 0.1 plus 0.2 doesn't equal exactly 0.3 in most programming languages. If you're building a system where exact decimal-to-fraction conversion matters, use a decimal data type instead of a float. It's a small change that prevents a class of bugs you won't notice until something goes wrong in production.
