Report a bug
If you spot a problem with this page, click here to create a Bugzilla issue.
Improve this page
Quickly fork, edit online, and submit a pull request for this page. Requires a signed-in GitHub account. This works well for small changes. If you'd like to make larger changes you may want to consider using a local clone.

std.functional

Functions that manipulate other functions.
This module provides functions for compile time function composition. These functions are helpful when constructing predicates for the algorithms in std.algorithm or std.range.
Function Name Description
adjoin Joins a couple of functions into one that executes the original functions independently and returns a tuple with all the results.
compose, pipe Join a couple of functions into one that executes the original functions one after the other, using one function's result for the next function's argument.
forward Forwards function arguments while saving ref-ness.
lessThan, greaterThan, equalTo Ready-made predicate functions to compare two values.
memoize Creates a function that caches its result for fast re-evaluation.
not Creates a function that negates another.
partial Creates a function that binds the first argument of a given function to a given value.
curry Converts a multi-argument function into a series of single-argument functions. f(x, y) == curry(f)(x)(y)
reverseArgs Predicate that reverses the order of its arguments.
toDelegate Converts a callable to a delegate.
unaryFun, binaryFun Create a unary or binary function from a string. Most often used when defining algorithms on ranges.
template unaryFun(alias fun, string parmName = "a")
Transforms a string representing an expression into a unary function. The string must either use symbol name a as the parameter or provide the symbol via the parmName argument.
Parameters:
fun a string or a callable
parmName the name of the parameter if fun is a string. Defaults to "a".
Returns:
If fun is a string, a new single parameter function
If fun is not a string, an alias to fun.
Examples:
// Strings are compiled into functions:
alias isEven = unaryFun!("(a & 1) == 0");
assert(isEven(2) && !isEven(1));
template binaryFun(alias fun, string parm1Name = "a", string parm2Name = "b")
Transforms a string representing an expression into a binary function. The string must either use symbol names a and b as the parameters or provide the symbols via the parm1Name and parm2Name arguments.
Parameters:
fun a string or a callable
parm1Name the name of the first parameter if fun is a string. Defaults to "a".
parm2Name the name of the second parameter if fun is a string. Defaults to "b".
Returns:
If fun is not a string, binaryFun aliases itself away to fun.
Examples:
alias less = binaryFun!("a < b");
assert(less(1, 2) && !less(2, 1));
alias greater = binaryFun!("a > b");
assert(!greater("1", "2") && greater("2", "1"));
alias lessThan = safeOp!"<".safeOp(T0, T1)(auto ref T0 a, auto ref T1 b);
Predicate that returns a < b. Correctly compares signed and unsigned integers, ie. -1 < 2U.
Examples:
assert(lessThan(2, 3));
assert(lessThan(2U, 3U));
assert(lessThan(2, 3.0));
assert(lessThan(-2, 3U));
assert(lessThan(2, 3U));
assert(!lessThan(3U, -2));
assert(!lessThan(3U, 2));
assert(!lessThan(0, 0));
assert(!lessThan(0U, 0));
assert(!lessThan(0, 0U));
alias greaterThan = safeOp!">".safeOp(T0, T1)(auto ref T0 a, auto ref T1 b);
Predicate that returns a > b. Correctly compares signed and unsigned integers, ie. 2U > -1.
Examples:
assert(!greaterThan(2, 3));
assert(!greaterThan(2U, 3U));
assert(!greaterThan(2, 3.0));
assert(!greaterThan(-2, 3U));
assert(!greaterThan(2, 3U));
assert(greaterThan(3U, -2));
assert(greaterThan(3U, 2));
assert(!greaterThan(0, 0));
assert(!greaterThan(0U, 0));
assert(!greaterThan(0, 0U));
alias equalTo = safeOp!"==".safeOp(T0, T1)(auto ref T0 a, auto ref T1 b);
Predicate that returns a == b. Correctly compares signed and unsigned integers, ie. !(-1 == ~0U).
Examples:
assert(equalTo(0U, 0));
assert(equalTo(0, 0U));
assert(!equalTo(-1, ~0U));
template reverseArgs(alias pred)
N-ary predicate that reverses the order of arguments, e.g., given pred(a, b, c), returns pred(c, b, a).
Parameters:
pred A callable
Returns:
A function which calls pred after reversing the given parameters
Examples:
alias gt = reverseArgs!(binaryFun!("a < b"));
assert(gt(2, 1) && !gt(1, 1));
Examples:
int x = 42;
bool xyz(int a, int b) { return a * x < b / x; }
auto foo = &xyz;
foo(4, 5);
alias zyx = reverseArgs!(foo);
writeln(zyx(5, 4)); // foo(4, 5)
Examples:
alias gt = reverseArgs!(binaryFun!("a < b"));
assert(gt(2, 1) && !gt(1, 1));
int x = 42;
bool xyz(int a, int b) { return a * x < b / x; }
auto foo = &xyz;
foo(4, 5);
alias zyx = reverseArgs!(foo);
writeln(zyx(5, 4)); // foo(4, 5)
Examples:
int abc(int a, int b, int c) { return a * b + c; }
alias cba = reverseArgs!abc;
writeln(abc(91, 17, 32)); // cba(32, 17, 91)
Examples:
int a(int a) { return a * 2; }
alias _a = reverseArgs!a;
writeln(a(2)); // _a(2)
Examples:
int b() { return 4; }
alias _b = reverseArgs!b;
writeln(b()); // _b()
template not(alias pred)
Negates predicate pred.
Parameters:
pred A string or a callable
Returns:
A function which calls pred and returns the logical negation of its return value.
Examples:
import std.algorithm.searching : find;
import std.functional;
import std.uni : isWhite;
string a = "   Hello, world!";
writeln(find!(not!isWhite)(a)); // "Hello, world!"
template partial(alias fun, alias arg)
Partially applies fun by tying its first argument to arg.
Parameters:
fun A callable
arg The first argument to apply to fun
Returns:
A new function which calls fun with arg plus the passed parameters.
Examples:
int fun(int a, int b) { return a + b; }
alias fun5 = partial!(fun, 5);
writeln(fun5(6)); // 11
// Note that in most cases you'd use an alias instead of a value
// assignment. Using an alias allows you to partially evaluate template
// functions without committing to a particular type of the function.
auto curry(alias F)()
if (isCallable!F && Parameters!F.length);

auto curry(T)(T t)
if (isCallable!T && Parameters!T.length);
Takes a function of (potentially) many arguments, and returns a function taking one argument and returns a callable taking the rest. f(x, y) == curry(f)(x)(y)
Parameters:
F a function taking at least one argument
T t a callable object whose opCall takes at least 1 object
Returns:
A single parameter callable object
Examples:
int f(int x, int y, int z)
{
    return x + y + z;
}
auto cf = curry!f;
auto cf1 = cf(1);
auto cf2 = cf(2);

writeln(cf1(2)(3)); // f(1, 2, 3)
writeln(cf2(2)(3)); // f(2, 2, 3)
Examples:
//works with callable structs too
struct S
{
    int w;
    int opCall(int x, int y, int z)
    {
        return w + x + y + z;
    }
}

S s;
s.w = 5;

auto cs = curry(s);
auto cs1 = cs(1);
auto cs2 = cs(2);

writeln(cs1(2)(3)); // s(1, 2, 3)
writeln(cs1(2)(3)); // (1 + 2 + 3 + 5)
writeln(cs2(2)(3)); // s(2, 2, 3)
template adjoin(F...) if (F.length == 1)

template adjoin(F...) if (F.length > 1)
Takes multiple functions and adjoins them together.
Parameters:
F the call-able(s) to adjoin
Returns:
A new function which returns a std.typecons.Tuple. Each of the elements of the tuple will be the return values of F.

Note In the special case where only a single function is provided (F.length == 1), adjoin simply aliases to the single passed function (F[0]).

Examples:
import std.functional, std.typecons : Tuple;
static bool f1(int a) { return a != 0; }
static int f2(int a) { return a / 2; }
auto x = adjoin!(f1, f2)(5);
assert(is(typeof(x) == Tuple!(bool, int)));
assert(x[0] == true && x[1] == 2);
template compose(fun...)
Composes passed-in functions fun[0], fun[1], ....
Parameters:
fun the call-able(s) or string(s) to compose into one function
Returns:
A new function f(x) that in turn returns fun[0](fun[1](...(x)))....
See Also:
Examples:
import std.algorithm.comparison : equal;
import std.algorithm.iteration : map;
import std.array : split;
import std.conv : to;

// First split a string in whitespace-separated tokens and then
// convert each token into an integer
assert(compose!(map!(to!(int)), split)("1 2 3").equal([1, 2, 3]));
template pipe(fun...)
Pipes functions in sequence. Offers the same functionality as compose, but with functions specified in reverse order. This may lead to more readable code in some situation because the order of execution is the same as lexical order.
Parameters:
fun the call-able(s) or string(s) to compose into one function
Returns:
A new function f(x) that in turn returns fun[$-1](...fun[1](fun[0](x)))....

Example

// Read an entire text file, split the resulting string in
// whitespace-separated tokens, and then convert each token into an
// integer
int[] a = pipe!(readText, split, map!(to!(int)))("file.txt");

See Also:
Examples:
import std.conv : to;
string foo(int a) { return to!(string)(a); }
int bar(string a) { return to!(int)(a) + 1; }
double baz(int a) { return a + 0.5; }
writeln(compose!(baz, bar, foo)(1)); // 2.5
writeln(pipe!(foo, bar, baz)(1)); // 2.5

writeln(compose!(baz, `to!(int)(a) + 1`, foo)(1)); // 2.5
writeln(compose!(baz, bar)("1"[])); // 2.5

writeln(compose!(baz, bar)("1")); // 2.5

writeln(compose!(`a + 0.5`, `to!(int)(a) + 1`, foo)(1)); // 2.5
ReturnType!fun memoize(alias fun)(Parameters!fun args);

ReturnType!fun memoize(alias fun, uint maxSize)(Parameters!fun args);
Memoizes a function so as to avoid repeated computation. The memoization structure is a hash table keyed by a tuple of the function's arguments. There is a speed gain if the function is repeatedly called with the same arguments and is more expensive than a hash table lookup. For more information on memoization, refer to this book chapter.

Example

double transmogrify(int a, string b)
{
   ... expensive computation ...
}
alias fastTransmogrify = memoize!transmogrify;
unittest
{
    auto slow = transmogrify(2, "hello");
    auto fast = fastTransmogrify(2, "hello");
    assert(slow == fast);
}

Parameters:
fun the call-able to memozie
maxSize The maximum size of the GC buffer to hold the return values
Returns:
A new function which calls fun and caches its return values.

Note Technically the memoized function should be pure because memoize assumes it will always return the same result for a given tuple of arguments. However, memoize does not enforce that because sometimes it is useful to memoize an impure function, too.

Examples:
To memoize a recursive function, simply insert the memoized call in lieu of the plain recursive call. For example, to transform the exponential-time Fibonacci implementation into a linear-time computation:
ulong fib(ulong n) @safe nothrow
{
    return n < 2 ? n : memoize!fib(n - 2) + memoize!fib(n - 1);
}
writeln(fib(10)); // 55
Examples:
To improve the speed of the factorial function,
ulong fact(ulong n) @safe
{
    return n < 2 ? 1 : n * memoize!fact(n - 1);
}
writeln(fact(10)); // 3628800
Examples:
This memoizes all values of fact up to the largest argument. To only cache the final result, move memoize outside the function as shown below.
ulong factImpl(ulong n) @safe
{
    return n < 2 ? 1 : n * factImpl(n - 1);
}
alias fact = memoize!factImpl;
writeln(fact(10)); // 3628800
Examples:
When the maxSize parameter is specified, memoize will used a fixed size hash table to limit the number of cached entries.
ulong fact(ulong n)
{
    // Memoize no more than 8 values
    return n < 2 ? 1 : n * memoize!(fact, 8)(n - 1);
}
writeln(fact(8)); // 40320
// using more entries than maxSize will overwrite existing entries
writeln(fact(10)); // 3628800
auto toDelegate(F)(auto ref F fp)
if (isCallable!F);
Convert a callable to a delegate with the same parameter list and return type, avoiding heap allocations and use of auxiliary storage.
Parameters:
F fp a function pointer or an aggregate type with opCall defined.
Returns:
A delegate with the context pointer pointing to nothing.

Example

void doStuff() {
    writeln("Hello, world.");
}

void runDelegate(void delegate() myDelegate) {
    myDelegate();
}

auto delegateToPass = toDelegate(&doStuff);
runDelegate(delegateToPass);  // Calls doStuff, prints "Hello, world."

Bugs:
  • Does not work with @safe functions.
  • Ignores C-style / D-style variadic arguments.
Examples:
static int inc(ref uint num) {
    num++;
    return 8675309;
}

uint myNum = 0;
auto incMyNumDel = toDelegate(&inc);
auto returnVal = incMyNumDel(myNum);
writeln(myNum); // 1
template forward(args...)
Forwards function arguments while keeping out, ref, and lazy on the parameters.
Parameters:
args a parameter list or an std.meta.AliasSeq.
Returns:
An AliasSeq of args with out, ref, and lazy saved.
Examples:
class C
{
    static int foo(int n) { return 1; }
    static int foo(ref int n) { return 2; }
}

// with forward
int bar()(auto ref int x) { return C.foo(forward!x); }

// without forward
int baz()(auto ref int x) { return C.foo(x); }

int i;
writeln(bar(1)); // 1
writeln(bar(i)); // 2

writeln(baz(1)); // 2
writeln(baz(i)); // 2
Examples:
void foo(int n, ref string s) { s = null; foreach (i; 0 .. n) s ~= "Hello"; }

// forwards all arguments which are bound to parameter tuple
void bar(Args...)(auto ref Args args) { return foo(forward!args); }

// forwards all arguments with swapping order
void baz(Args...)(auto ref Args args) { return foo(forward!args[$/2..$], forward!args[0..$/2]); }

string s;
bar(1, s);
writeln(s); // "Hello"
baz(s, 2);
writeln(s); // "HelloHello"
Examples:
struct X {
    int i;
    this(this)
    {
        ++i;
    }
}

struct Y
{
    private X x_;
    this()(auto ref X x)
    {
        x_ = forward!x;
    }
}

struct Z
{
    private const X x_;
    this()(auto ref X x)
    {
        x_ = forward!x;
    }
    this()(auto const ref X x)
    {
        x_ = forward!x;
    }
}

X x;
const X cx;
auto constX = (){ const X x; return x; };
static assert(__traits(compiles, { Y y = x; }));
static assert(__traits(compiles, { Y y = X(); }));
static assert(!__traits(compiles, { Y y = cx; }));
static assert(!__traits(compiles, { Y y = constX(); }));
static assert(__traits(compiles, { Z z = x; }));
static assert(__traits(compiles, { Z z = X(); }));
static assert(__traits(compiles, { Z z = cx; }));
static assert(__traits(compiles, { Z z = constX(); }));


Y y1 = x;
// ref lvalue, copy
writeln(y1.x_.i); // 1
Y y2 = X();
// rvalue, move
writeln(y2.x_.i); // 0

Z z1 = x;
// ref lvalue, copy
writeln(z1.x_.i); // 1
Z z2 = X();
// rvalue, move
writeln(z2.x_.i); // 0
Z z3 = cx;
// ref const lvalue, copy
writeln(z3.x_.i); // 1
Z z4 = constX();
// const rvalue, copy
writeln(z4.x_.i); // 1