Literals

This article provides a table that shows how to specify the type of a literal in F#.

Literal types

The following table shows the literal types in F#. Characters that represent digits in hexadecimal notation are not case-sensitive; characters that identify the type are case-sensitive.

Named literals

Values that are intended to be constants can be marked with the Literal attribute.

This attribute has the effect of causing a value to be compiled as a constant. In the following example, both x and y below are immutable values, but x is evaluated at run-time, whereas y is a compile-time constant.

let x = "a" + "b" // evaluated at run-time

[<Literal>]
let y = "a" + "b" // evaluated at compile-time

For example, this distinction matters when calling an external function, because DllImport is an attribute that needs to know the value of myDLL during compilation. Without the [<Literal>] declaration, this code would fail to compile:

[<Literal>]
let myDLL = "foo.dll"

[<DllImport(myDLL, CallingConvention = CallingConvention.Cdecl)>]
extern void HelloWorld()

In pattern matching expressions, identifiers that begin with lowercase characters are always treated as variables to be bound, rather than as literals, so you should generally use initial capitals when you define literals.

[<Literal>]
let SomeJson = """{"numbers":[1,2,3,4,5]}"""

[<Literal>]
let Literal1 = "a" + "b"

[<Literal>]
let FileLocation =   __SOURCE_DIRECTORY__ + "/" + __SOURCE_FILE__

[<Literal>]
let Literal2 = 1 ||| 64

[<Literal>]
let Literal3 = System.IO.FileAccess.Read ||| System.IO.FileAccess.Write

Remarks

Named literals are useful for:

• Pattern matching without a when clause.
• Attribute arguments.
• Static type provider arguments.

Unicode strings can contain explicit encodings that you can specify by using \u followed by a 16-bit hexadecimal code (0000 - FFFF), or UTF-32 encodings that you can specify by using \U followed by a 32-bit hexadecimal code that represents any Unicode code point (00000000 - 0010FFFF).

The use of bitwise operators other than ||| isn't allowed.

Integers in other bases

Signed 32-bit integers can also be specified in hexadecimal, octal, or binary using a 0x, 0o or 0b prefix, respectively.

let numbers = (0x9F, 0o77, 0b1010)
// Result: numbers : int * int * int = (159, 63, 10)

Underscores in numeric literals

You can separate digits with the underscore character (_).

let value = 0xDEAD_BEEF

let valueAsBits = 0b1101_1110_1010_1101_1011_1110_1110_1111

let exampleSSN = 123_45_6789

Special floating-point infinity values

Both the float and single floating-point numeric types have associated special values representing positive and negative infinity.

These values can be used directly or are returned when dividing by a floating-point zero or a number too small to be represented by the given type. For example:

> 1.0/0.0;;
val it: float = infinity

> 1.0/(-0.0);;
val it: float = -infinity

> 1.0/0.00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000001
;;
val it: float = infinity