Description

The following code:

<?php
$units = ['5', '10', '5', '3A', '5', '5'];
$unique = array_unique($units, SORT_REGULAR);
print_r($unique);

Resulted in this output:

Array
(
    [0] => 5
    [1] => 10
    [3] => 3A
    [4] => 5
)

But I expected this output instead:

Array
(
    [0] => 5
    [1] => 10
    [3] => 3A
)

Demonstrations:

• simple array of strings: https://3v4l.org/M5lcP
• array of objects: https://3v4l.org/5kr1t
• array of arrays: https://3v4l.org/0HVoV

PRs in progress:

• Fix GH-20262: array_unique() SORT_REGULAR fails to deduplicate with mixed strings #20273 :
  This fixes the issue with an array of scalars and provides a performance boost to boot.
• Fix #20262: SORT_REGULAR transitivity violation with mixed numeric/non-numeric strings #20305 :
  This PR is an attempt to get to the core of the issue, the transitivity violation, which conveniently resolves the knock-on effect SORT_REGULAR has on other functions like sort() (see https://3v4l.org/lYP9Q).

Root Cause

The algorithm:

1. Sort array using comparison function from php_get_data_compare_func_unstable()
2. Walk through sorted array comparing only adjacent elements
3. Delete duplicates when adjacent elements compare equal

The bug:

SORT_REGULAR uses zend_compare() which calls zendi_smart_strcmp() for string comparisons. This function has non-transitive behavior when mixing numeric and non-numeric strings:

• "5" < "10" → true (numeric comparison: 5 < 10)
• "10" < "3A" → true (lexicographic: "1" < "3")
• "3A" < "5" → true (lexicographic: "3" < "5") Creates a cycle!

Because the comparison is non-transitive, sorting algorithms (which require transitive comparisons) produce inconsistent results depending on input order.

The deduplication walks through comparing adjacent elements:

lastkept = position_0;  // "5"
position_1 "10" != "5" → keep, lastkept = position_1
position_2 "10" == "10" → delete
position_3 "3A" != "10" → keep, lastkept = position_3
position_4 "5" != "3A" → keep  ← Bug! Never compared to position_0
position_5 "5" == "5" → delete

The root issue: Non-transitive comparisons break the sorting algorithm's guarantee that equal values will be grouped together. The adjacent-only comparison is correct - but it requires the array to be properly sorted first, which requires transitive comparisons.

Comparison with SORT_STRING

<?php
$units = ['5', '10', '5', '3A', '5', '5'];
echo count(array_unique($units, SORT_REGULAR)) . "\n"; // 4 ✗ Wrong
echo count(array_unique($units, SORT_STRING)) . "\n";  // 3 ✓ Correct

SORT_STRING uses lexical comparison without numeric extraction, so duplicates stay grouped.

Workaround

For simple arrays of scalar values, you can use array_unique with default SORT_STRING flag.

<?php
$unique = array_unique($array, SORT_STRING);

For arrays or objects.

$uniqueAddr = [];
foreach ($addresses as $addr) {
    if (! in_array($addr, $uniqueAddr)) {
        $uniqueAddr[] = $addr;
    }
}

PHP Version

PHP 8.4.13 (cli) (built: Sep 26 2025 00:45:36) (NTS clang 15.0.0)
Copyright (c) The PHP Group
Built by Laravel Herd
Zend Engine v4.4.13, Copyright (c) Zend Technologies
    with Zend OPcache v8.4.13, Copyright (c), by Zend Technologies