I find a very simple solution to count values in multidimentional arrays (example for 2 levels) :
foreach ($array as $a) {
foreach ($a as $b) {
$count_values[$b]++;
}
}
(PHP 4, PHP 5)
array_count_values — Zählt die Werte eines Arrays
$array
)
array_count_values() gibt ein Array zurück, in
dem die Werte des Arrays array
als
Schlüssel, und die Häufigkeit ihres Auftretens als Werte angegeben sind.
array
Das Array mit den zu zählenden Werten
Gibt ein assoziatives Array mit den Werten aus array
als Schlüsseln und der Anzahl als Werten zurück.
Beispiel #1 array_count_values()-Beispiel
<?php
$array = array(1, "hello", 1, "world", "hello");
print_r(array_count_values($array));
?>
Das oben gezeigte Beispiel erzeugt folgende Ausgabe:
Array ( [1] => 2 [hello] => 2 [world] => 1 )
I find a very simple solution to count values in multidimentional arrays (example for 2 levels) :
foreach ($array as $a) {
foreach ($a as $b) {
$count_values[$b]++;
}
}
In order to apply array_map with callback checking for localised values like city name, country name, you have to provide some sort of comparison array.
Here's an example of array_count_values for Polish city names. Just switch array_keys / array_values in order to obtain lowercase / uppercase.
$arr = array('Warsaw', 'Warsaw', 'Wrocław', 'Poznań', 'KrakÓw', 'waRsaw', 'Gdańsk', 'poznań', 'WROCŁAW', 'kraków', 'GDAŃSK');
$lowertoupperpolish = array(
'ą' => 'Ą',
'ć' => 'Ć',
'ę' => 'Ę',
'ł' => 'Ł',
'ń' => 'Ń',
'ó' => 'Ó',
'ś' => 'Ś',
'ż' => 'Ż',
'ź' => 'Ź',
);
function lowers($n) {
global $lowertoupperpolish;
return strtolower(str_replace(array_values($lowertoupperpolish), array_keys($lowertoupperpolish), $n));
}
$result = array_count_values(array_map('lowers', $arr));
print_r($result); ...prints:
Array ( [warsaw] => 3 [wrocław] => 2 [poznań] => 2 [kraków] => 2 [gdańsk] => 2 )
<?
function array_icount_values($array) {
$ret_array = array();
foreach($array as $value) $ret_array[strtolower($value)]++;
return $ret_array;
}
$ar = array('J. Karjalainen', 'J. Karjalainen', 60, '60', 'J. Karjalainen', 'j. karjalainen', 'Fastway', 'FASTWAY', 'Fastway', 'fastway', 'YUP');
$ar = array_icount_values($ar);
?>
this prints:
Array
(
[j. karjalainen] => 4
[60] => 2
[fastway] => 4
[yup] => 1
)
Yet Another case-insensitive version of array_count_values()
<?php
$ar = array('J. Karjalainen', 'J. Karjalainen', 60, '60', 'J. Karjalainen', 'j. karjalainen', 'Fastway', 'FASTWAY', 'Fastway', 'fastway', 'YUP');
$ar = array_count_values(array_map('strtolower', $ar));
?>
The case-insensitive version:
<?php
function array_count_values_ci($array) {
$newArray = array();
foreach ($array as $values) {
if (!array_key_exists(strtolower($values), $newArray)) {
$newArray[strtolower($values)] = 0;
}
$newArray[strtolower($values)] += 1;
}
return $newArray;
}
?>
I couldn't find a function for counting the values with case-insensitive matching, so I wrote a quick and dirty solution myself:
<pre><?php
function array_icount_values($array) {
$ret_array = array();
foreach($array as $value) {
foreach($ret_array as $key2 => $value2) {
if(strtolower($key2) == strtolower($value)) {
$ret_array[$key2]++;
continue 2;
}
}
$ret_array[$value] = 1;
}
return $ret_array;
}
$ar = array('J. Karjalainen', 'J. Karjalainen', 60, '60', 'J. Karjalainen', 'j. karjalainen', 'Fastway', 'FASTWAY', 'Fastway', 'fastway', 'YUP');
$ar2 = array_count_values($ar); // Normal matching
$ar = array_icount_values($ar); // Case-insensitive matching
print_r($ar2);
print_r($ar);
?></pre>
This prints:
Array
(
[J. Karjalainen] => 3
[60] => 2
[j. karjalainen] => 1
[Fastway] => 2
[FASTWAY] => 1
[fastway] => 1
[YUP] => 1
)
Array
(
[J. Karjalainen] => 4
[60] => 2
[Fastway] => 4
[YUP] => 1
)
I don't know how efficient it is, but it seems to work. Needed this function in one of my scripts and thought I would share it.
Here is a Version with one or more arrays, which have similar values in it:
Use $lower=true/false to ignore/set case Sensitiv.
<?php
$ar1[] = array("red","green","yellow","blue");
$ar1[] = array("green","yellow","brown","red","white","yellow");
$ar1[] = array("red","green","brown","blue","black","yellow");
#$ar1= array("red","green","brown","blue","black","red","green"); // Possible with one or multiple Array
$res = array_icount_values ($ar1);
print_r($res);
function array_icount_values($arr,$lower=true) {
$arr2=array();
if(!is_array($arr['0'])){$arr=array($arr);}
foreach($arr as $k=> $v){
foreach($v as $v2){
if($lower==true) {$v2=strtolower($v2);}
if(!isset($arr2[$v2])){
$arr2[$v2]=1;
}else{
$arr2[$v2]++;
}
}
}
return $arr2;
}
/*
Will print:
Array
(
[red] => 3
[green] => 3
[yellow] => 4
[blue] => 2
[brown] => 2
[white] => 1
[black] => 1
)
*/
?>
array_count_values function does not work on multidimentional arrays.
If $score[][] is a bidimentional array, the command
"array_count_values ($score)" return the error message "Warning: Can only count STRING and INTEGER values!".
You might use serialize() to serialize your objects before analyzing their frequency. :)
byron at byronrode dot co dot za, here are some benchmarks.
<?php
$haystack = Array();
for ($i = 0; $i < 1000000; $i++) {
$haystack[] = rand(1, 2000);
}
$needle = rand(1, 2000);
echo "__array_count_values()__\n";
$start = microtime(true);
$startmem = memory_get_usage();
$counts = array_count_values($haystack);
$mem = memory_get_usage()-$startmem;
echo 'Count:'.$counts[$needle]."\n";
echo 'Time:'.(microtime(true) - $start)."\n";
echo 'Memory:'.$mem."\n\n";
echo "__array_keys()__\n";
$start = microtime(true);
$startmem = memory_get_usage();
$keys = array_keys($haystack, $needle);
$mem = memory_get_usage()-$startmem;
echo 'Count:'.count($keys)."\n";
echo 'Time:'.(microtime(true) - $start)."\n";
echo 'Memory:'.$mem."\n\n";
echo '__$needle_array[]__'."\n";
$start = microtime(true);
$startmem = memory_get_usage();
$x = count($haystack);
for($i = 0; $i < $x; $i++){
if($haystack[$i] == $needle){
$needle_array[] = $haystack[$i];
}
}
$mem = memory_get_usage()-$startmem;
$number_of_instances = count($needle_array);
echo 'Count:'.$number_of_instances."\n";
echo 'Time:'.(microtime(true) - $start)."\n";
echo 'Memory:'.$mem."\n\n";
echo '__$number_of_instances++__'."\n";
$start = microtime(true);
$startmem = memory_get_usage();
$x = count($haystack);
$number_of_instances = 0;
for($i = 0; $i < $x; $i++){
if($haystack[$i] == $needle){
$number_of_instances++;
}
}
$mem = memory_get_usage()-$startmem;
echo 'Count:'.$number_of_instances."\n";
echo 'Time:'.(microtime(true) - $start)."\n";
echo 'Memory:'.$mem."\n\n";
?>
[www]mytemp$ php array_count_test.php
__array_count_values()__
Count:515
Time:0.0607650279999
Memory:120328
__array_keys()__
Count:515
Time:0.0869138240814
Memory:33016
__$needle_array[]__
Count:515
Time:0.259949922562
Memory:24792
__$number_of_instances++__
Count:515
Time:0.258481025696
Memory:0
However, when you use an array of strings by calling md5(rand(1, 2000)), the performance boosts become less significant:
__array_count_values()__
Count:499
Time:0.491794109344
Memory:184328
__array_keys()__
Count:499
Time:0.36399102211
Memory:30072
__$needle_array[]__
Count:499
Time:0.568728923798
Memory:22104
__$number_of_instances++__
Count:499
Time:0.574353933334
Memory:0
Results are similar for string->string haystacks with foreach traversal.
alwaysdrunk's comment only works if you can trust the client web browser. Using this function doesn't validate that every necessary field exists -- only that every field that was submitted has a value in it. Thus if an attacker wished to force a null value into one of the fields, he could (rather easily) construct a modified form without the field and submit THAT.
Besides, you really ought to be validating each field anyway if you're taking user input.