Floating Point Numbers in

About Floating Point Numbers

A floating-point number is a number with zero or more digits behind the decimal separator. Examples are -2.4, 0.1, 3.14, 16.984025 and 1024.0.

F# has three floating-point types:

• single: 4 bytes (~6-9 digits precision). Written as 2.45f or 2.45F. The float32 type can be used as an alias.
• float: 8 bytes (~15-17 digits precision). This is the most common type. Written as 2.45 or using exponent notation as 2.3E+32 or 2.3e+32. The double type can be used as an alias.
• decimal: 16 bytes (28-29 digits precision). Normally used when working with monetary data, as its precision leads to less rounding errors. Written as 2.45m or 2.45M.

Each floating-point type has its own precision, approximate range and size. The precision indicates how many digits after the digit separator can be stored. This means that trying to store PI in a single will only store the first 6 to 9 digits (with the last digit being rounded).

Converting between different floating-point types is done using the conversion operators:

let floatFromSingle = float 2.45f

Always be careful when checking the values of floating-point types for equality, as values that can appear to represent the same value could actually be different. See this article for more information (the code examples are in C#).

You can find a short introduction to floating-point numbers at 0.30000000000000004.com. The Float Toy page has a nice, graphical explanation how a floating-point numbers' bits are converted to an actual floating-point value.