Associative arrays/Iteration
From Rosetta Code
Show how to iterate over the key-value pairs of an associative array, and print each pair out. Also show how to iterate just over the keys, or the values, if there is a separate way to do that in your language.
Contents |
[edit] Ada
with Ada.Text_IO; use Ada.Text_IO;
with Ada.Containers.Indefinite_Ordered_Maps;
procedure Test_Iteration is
package String_Maps is
new Ada.Containers.Indefinite_Ordered_Maps (String, Integer);
use String_Maps;
A : Map;
Index : Cursor;
begin
A.Insert ("hello", 1);
A.Insert ("world", 2);
A.Insert ("!", 3);
Index := A.First;
while Index /= No_Element loop
Put_Line (Key (Index) & Integer'Image (Element (Index)));
Index := Next (Index);
end loop;
end Test_Iteration;
Sample output:
! 3 hello 1 world 2
[edit] AWK
In AWK "arrays" are always associative arrays, and the only way to iterate over them is by keys (indexes in the AWK terminology)
BEGIN {
a["hello"] = 1
a["world"] = 2
a["!"] = 3
# iterate over keys
for(key in a) {
print key, a[key]
}
}
[edit] C
Library: Judy
We can easily iterate over pair of keys (indexes) and values.
#include <stdio.h>
#include <Judy.h>
#define MAXLINELEN 256
int main()
{
Pvoid_t assoc_arr = (Pvoid_t) NULL;
PWord_t val;
uint8_t idx[MAXLINELEN];
// insert some values
JSLI(val, assoc_arr, "hello");
*val = 4;
JSLI(val, assoc_arr, "world");
*val = 8;
JSLI(val, assoc_arr, "!");
*val = 16;
// iterate over indexes-values
idx[0] = '\0';
JSLF(val, assoc_arr, idx);
while(val != NULL) {
printf("'%s' -> %d\n", idx, *val);
JSLN(val, assoc_arr, idx);
}
JudySLFreeArray(&assoc_arr, PJE0);
return 0;
}
[edit] C++
std::map<std::string, int> myDict;
myDict["hello"] = 1;
myDict["world"] = 2;
myDict["!"] = 3;
// iterating over key-value pairs:
for (std::map<std::string, int>::iterator it = myDict.begin(); it != myDict.end(); it++) {
// the thing pointed to by the iterator is a pair<std::string, int>
std::string key = it->first;
int value = it->second;
std::cout << "key = " << key << ", value = " << value << std::endl;
}
[edit] Clojure
(doseq [[k v] {:a 1, :b 2, :c 3}]
(println k "=" v))
(doseq [k (keys {:a 1, :b 2, :c 3})]
(println k))
(doseq [v (vals {:a 1, :b 2, :c 3})]
(println v))
[edit] Common Lisp
Common Lisp has three common idioms for associating keys with values: association lists (alists), property lists (plists), and hash tables.
[edit] With association lists (alists)
The association list is a list of conses, each of whose car is a key and whose cdr is a value. The standard mapping and print functions can be used to print key/value pairs, keys, and values.
;; iterate using dolist, destructure manually
(dolist (pair alist)
(destructuring-bind (key . value) pair
(format t "~&Key: ~a, Value: ~a." key value)))
;; iterate and destructure with loop
(loop for (key . value) in alist
do (format t "~&Key: ~a, Value: ~a." key value))
[edit] With property lists (plists)
Property lists are lists of alternating keys and values, where each value's key is the element of the list immediately following it. Printing could be done with standard mapping functions, but loop's destructuring makes things a bit easier.
(loop for (key value) on plist :by 'cddr
do (format t "~&Key: ~a, Value: ~a." key value))
[edit] With hash tables
Lisp also has built-in hash tables, and there are several ways to map over these. The first is maphash which takes a function of two arguments (the key and value) and the hash table.
(maphash (lambda (key value)
(format t "~&Key: ~a, Value: ~a." key value))
hash-table)
The loop construct also supports extracting key/value pairs from hash tables.
(loop for key being each hash-key of hash-table using (hash-value value)
do (format t "~&Key: ~a, Value: ~a." key value))
There is also a macro with-hash-table-iterator which locally binds a name to produce associated keys and values of the hash table; while rarely used, it is the most powerful operation.
(with-hash-table-iterator (next-entry hash-table)
(loop
(multiple-value-bind (nextp key value) (next-entry)
(if (not nextp)
(return)
(format t "~&Key: ~a, Value: ~a." key value)))))
[edit] D
D V.2 code.
import std.stdio: writeln;
void main() {
// the associative array
int[string] aa = ["alice":2, "bob":97, "charlie":45];
// how to extract key/value pairs:
foreach (key, value; aa)
writeln("Got key ", key, " with value ", value);
// how to extract just the keys:
foreach (key; aa.keys)
writeln("Got key ", key);
// how to extract just the values:
foreach (value; aa)
writeln("Got value ",value);
// how to extract just the values, alternative:
foreach (value; aa.values)
writeln("Got value ",value);
}
[edit] E
In E, the basic iteration protocol and syntax work over key-value pairs. Therefore, any iteration over a map or other collection is always key-value, though the user may choose to ignore the keys or the values.
The for loop takes either one pattern, for the value, or two, for the key and value; for iterating over keys alone the value may be given an ignore-pattern (_).
def map := [
"a" => 1,
"b" => 2,
"c" => 3,
]
for key => value in map {
println(`$key $value`)
}
for value in map { # ignore keys
println(`. $value`)
}
for key => _ in map { # ignore values
println(`$key .`)
}
for key in map.domain() { # iterate over the set whose values are the keys
println(`$key .`)
}
[edit] Factor
H{ { "hi" "there" } { "a" "b" } } [ ": " glue print ] assoc-each
There's also assoc-map, assoc-find, assoc-filter and many more.
[edit] Haskell
with Data.Map:
import qualified Data.Map as M
myMap = M.fromList [("hello", 13), ("world", 31), ("!", 71)]
main = do -- pairs
print $ M.toList myMap
-- keys
print $ M.keys myMap
-- values
print $ M.elems myMap
[edit] J
Note that all J operations either iterate over the items of an array or can be made to do so. So to iterate over some sequence you need to refer to that sequence.
Using the J example from Creating an Associative Array...
Keysnl__example 0
Values get__example each nl__example 0
Both keys and values (,&< get__example) each nl__example 0
Note that this last is not likely to be useful in any practical context outside of learning the language.
[edit] Java
Map<String, Integer> myDict = new HashMap<String, Integer>();
myDict.put("hello", 1);
myDict.put("world", 2);
myDict.put("!", 3);
// iterating over key-value pairs:
for (Map.Entry<String, Integer> e : myDict.entrySet()) {
String key = e.getKey();
Integer value = e.getValue();
System.out.println("key = " + key + ", value = " + value);
}
// iterating over keys:
for (String key : myDict.keySet()) {
System.out.println("key = " + key);
}
// iterating over values:
for (Integer value : myDict.values()) {
System.out.println("value = " + value);
}
[edit] JavaScript
JavaScript does not have associative arrays. You can add properties to an empty object, and that works basically the same way:
var myhash = {}; // a new, empty object
myhash["hello"] = 3;
myhash.world = 6; // obj.name is equivalent to obj["name"] for certain values of name
myhash["!"] = 9;
var output;
for (var key in myhash) {
output += "key is: " + key;
output += " => ";
output += "value is: " + myhash[key]; // cannot use myhash.key, that would be myhash["key"]
output += "\n";
}
[edit] Lua
local table = {
["foo"] = "bar",
["baz"] = 6,
42 = 7,
}
for key,val in pairs(table) do
print(string.format("%s: %s\n", key, val)
end
[edit] M4
divert(-1)
define(`for',
`ifelse($#,0,``$0'',
`ifelse(eval($2<=$3),1,
`pushdef(`$1',$2)$4`'popdef(`$1')$0(`$1',incr($2),$3,`$4')')')')
define(`new',`define(`$1[size]key',0)')
define(`asize',`defn(`$1[size]key')')
define(`aget',`defn(`$1[$2]')')
define(`akget',`defn(`$1[$2]key')')
define(`avget',`aget($1,akget($1,$2))')
define(`aset',
`ifdef($1[$2],
`',
`define(`$1[size]key',incr(asize(`$1')))`'define($1[asize(`$1')]key,$2)')`'define($1[$2],$3)')
define(`dquote', ``$@'')
define(`akeyvalue',`dquote(akget($1,$2),aget($1,akget($1,$2)))')
define(`akey',`dquote(akget($1,$2))')
define(`avalue',`dquote(aget($1,akget($1,$2)))')
divert
new(`a')
aset(`a',`wow',5)
aset(`a',`wow',flame)
aset(`a',`bow',7)
key-value pairs
for(`x',1,asize(`a'),
`akeyvalue(`a',x)
')
keys
for(`x',1,asize(`a'),
`akey(`a',x)
')
values
for(`x',1,asize(`a'),
`avalue(`a',x)
')
Output:
key-value pairs `wow',`flame' `bow',`7' keys `wow' `bow' values `flame' `7'
[edit] Objective-C
Works with: Objective-C version 2.0+
NSDictionary *myDict = [NSDictionary dictionaryWithObjectsAndKeys:
[NSNumber numberWithInt:13], @"hello",
[NSNumber numberWithInt:31], @"world",
[NSNumber numberWithInt:71], @"!", nil];
// iterating over keys:
for (id key in myDict) {
NSLog(@"key = %@", key);
}
// iterating over values:
for (id value in [myDict objectEnumerator]) {
NSLog(@"value = %@", value);
}
Works with: Objective-C version <2.0
NSDictionary *myDict = [NSDictionary dictionaryWithObjectsAndKeys:
[NSNumber numberWithInt:13], @"hello",
[NSNumber numberWithInt:31], @"world",
[NSNumber numberWithInt:71], @"!", nil];
// iterating over keys:
NSEnumerator *enm = [myDict keyEnumerator];
id key;
while ((key = [enm nextObject])) {
NSLog(@"key = %@", key);
}
// iterating over values:
enm = [myDict objectEnumerator];
id value;
while ((value = [enm nextObject])) {
NSLog(@"value = %@", value);
}
[edit] OCaml
Association list:
#!/usr/bin/env ocaml
let map = [| ('A', 1); ('B', 2); ('C', 3) |] ;;
(* iterate over pairs *)
Array.iter (fun (k,v) -> Printf.printf "key: %c - value: %d\n" k v) map ;;
(* iterate over keys *)
Array.iter (fun (k,_) -> Printf.printf "key: %c\n" k) map ;;
(* iterate over values *)
Array.iter (fun (_,v) -> Printf.printf "value: %d\n" v) map ;;
(* in functional programming it is often more useful to fold over the elements *)
Array.fold_left (fun acc (k,v) -> acc ^ Printf.sprintf "key: %c - value: %d\n" k v) "Elements:\n" map ;;
Hash table:
let map = Hashtbl.create 42;;
Hashtbl.add map 'A' 1;;
Hashtbl.add map 'B' 2;;
Hashtbl.add map 'C' 3;;
(* iterate over pairs *)
Hashtbl.iter (fun k v -> Printf.printf "key: %c - value: %d\n" k v) map ;;
(* in functional programming it is often more useful to fold over the elements *)
Hashtbl.fold (fun k v acc -> acc ^ Printf.sprintf "key: %c - value: %d\n" k v) map "Elements:\n" ;;
Functional binary search tree:
module CharMap = Map.Make (Char);;
let map = CharMap.empty;;
let map = CharMap.add 'A' 1 map;;
let map = CharMap.add 'B' 2 map;;
let map = CharMap.add 'C' 3 map;;
(* iterate over pairs *)
CharMap.iter (fun k v -> Printf.printf "key: %c - value: %d\n" k v) map ;;
(* in functional programming it is often more useful to fold over the elements *)
CharMap.fold (fun k v acc -> acc ^ Printf.sprintf "key: %c - value: %d\n" k v) map "Elements:\n" ;;
[edit] Oz
declare
MyMap = unit('hello':13 'world':31 '!':71)
in
{ForAll {Record.toListInd MyMap} Show} %% pairs
{ForAll {Record.arity MyMap} Show} %% keys
{ForAll {Record.toList MyMap} Show} %% values
[edit] Perl
#! /usr/bin/perl
use strict;
my %pairs = ( "hello" => 13,
"world" => 31,
"!" => 71 );
# iterate over pairs
while ( my ($k, $v) = each %pairs) {
print "(k,v) = ($k, $v)\n";
}
# iterate over keys
foreach my $key ( keys %pairs ) {
print "key = $key, value = $pairs{$key}\n";
}
# iterate over values
foreach my $val ( values %pairs ) {
print "value = $val\n";
}
[edit] Perl 6
Works with: Rakudo version #21 "Seattle"
my %pairs = hello => 13, world => 31, '!' => 71;
for %pairs.kv -> $k, $v {
say "(k,v) = ($k, $v)";
}
say "key = $_" for %pairs.keys;
say "value = $_" for %pairs.values;
[edit] PHP
<?php
$pairs = array( "hello" => 1,
"world" => 2,
"!" => 3 );
// iterate over key-value pairs
foreach($pairs as $k => $v) {
echo "(k,v) = ($k, $v)\n";
}
// iterate over keys
foreach(array_keys($pairs) as $key) {
echo "key = $key, value = $pairs[$key]\n";
}
// iterate over values
foreach($pairs as $value) {
echo "values = $value\n";
}
?>
[edit] PicoLisp
[edit] Using properties
(put 'A 'foo 5)
(put 'A 'bar 10)
(put 'A 'baz 15)
: (getl 'A) # Get the whole property list
-> ((15 . baz) (10 . bar) (5 . foo))
: (mapcar cdr (getl 'A)) # Get all keys
-> (baz bar foo)
: (mapcar car (getl 'A)) # Get all values
-> (15 10 5)
[edit] Using an index tree
(idx 'A (def "foo" 5) T)
(idx 'A (def "bar" 10) T)
(idx 'A (def "baz" 15) T)
: A # Get the whole tree
-> ("foo" ("bar" NIL "baz"))
: (idx 'A) # Get all keys
-> ("bar" "baz" "foo")
: (mapcar val (idx 'A)) # Get all values
-> (10 15 5)
[edit] PureBasic
Hashes are a built-in type called Map in Purebasic.
NewMap dict.s()
dict("de") = "German"
dict("en") = "English"
dict("fr") = "French"
ForEach dict()
Debug MapKey(dict()) + ":" + dict()
Next
[edit] Python
myDict = { "hello": 13,
"world": 31,
"!" : 71 }
# iterating over key-value pairs:
for key, value in myDict.items():
print ("key = %s, value = %s" % (key, value))
# iterating over keys:
for key in myDict:
print ("key = %s" % key)
# (is a shortcut for:)
for key in myDict.keys():
print ("key = %s" % key)
# iterating over values:
for value in myDict.values():
print ("value = %s" % value)
[edit] PowerShell
Using the following hash table:
$h = @{ 'a' = 1; 'b' = 2; 'c' = 3 }
Iterating over the key/value pairs is slightly cumbersome as it requires an explicit call to GetEnumerator:
$h.GetEnumerator() | ForEach-Object { Write-Host Key: $_.Name, Value: $_.Value }
A foreach statement can also be used:
foreach ($e in $h.GetEnumerator()) {
Write-Host Key: $e.Name, Value: $e.Value
}
Iterating over the keys:
$h.Keys | ForEach-Object { Write-Host Key: $_ }
foreach ($k in $h.Keys) {
Write-Host Key: $k
}
Iterating over the values:
$h.Values | ForEach-Object { Write-Host Value: $_ }
foreach ($v in $h.Values) {
Write-Host Value: $v
}
[edit] R
R does not have a built-in concept of key-value pairs, however vectors can have named elements, which is close.
x <- c(hello=1, world=2, "!"=3)
print(x)
hello world !
1 2 3
print(names(x))
"hello" "world" "!"
print(unname(x))
1 2 3
[edit] Ruby
myDict = { "hello" => 13,
"world" => 31,
"!" => 71 }
# iterating over key-value pairs:
myDict.each {|key, value| puts "key = #{key}, value = #{value}"}
# or
myDict.each_pair {|key, value| puts "key = #{key}, value = #{value}"}
# iterating over keys:
myDict.each_key {|key| puts "key = #{key}"}
# iterating over values:
myDict.each_value {|value| puts "value =#{value}"}
[edit] Slate
In Slate, all associative mappings inherit from Mapping, so they all have the same protocol. Even Sequences obey it, in addition to their own protocol for collections with ordered integer-range keys.
define: #pairs -> ({'hello' -> 1. 'world' -> 2. '!' -> 3. 'another!' -> 3} as: Dictionary).
pairs keysAndValuesDo: [| :key :value |
inform: '(k, v) = (' ; key printString ; ', ' ; value printString ; ')'
].
pairs keysDo: [| :key |
inform: '(k, v) = (' ; key printString ; ', ' ; (pairs at: key) printString ; ')'
].
pairs do: [| :value |
inform: 'value = ' ; value printString
].
[edit] Smalltalk
Works with: GNU Smalltalk
|pairs|
pairs := Dictionary
from: { 'hello' -> 1. 'world' -> 2. '!' -> 3. 'another!' -> 3 }.
"iterate over keys and values"
pairs keysAndValuesDo: [ :k :v |
('(k, v) = (%1, %2)' % { k. v }) displayNl
].
"iterate over keys"
pairs keysDo: [ :key |
('key = %1, value = %2' % { key. pairs at: key }) displayNl
].
"iterate over values"
pairs do: [ :value |
('value = %1' % { value }) displayNl
].
We could also obtain a set of keys or a collection of values and iterate over them with "do:":
(pairs keys) do: [ :k | "..." ].
(pairs values) do: [ :v | "..." ].
[edit] Tcl
[edit] With Arrays
array set myAry {
# list items here...
}
# Iterate over keys and values
foreach {key value} [array get myAry] {
puts "$key -> $value"
}
# Iterate over just keys
foreach key [array names myAry] {
puts "key = $key"
}
# There is nothing for directly iterating over just the values
# Use the keys+values version and ignore the keys
[edit] With Dictionaries
Works with: Tcl version 8.5
set myDict [dict create ...]; # Make the dictionary
# Iterate over keys and values
dict for {key value} $myDict {
puts "$key -> $value"
}
# Iterate over keys
foreach key [dict keys $myDict] {
puts "key = $key"
}
# Iterate over values
foreach value [dict values $myDict] {
puts "value = $value"
}
