php detect and convert dates from a string

Question

I was wondering if I could somehow to detect a date in a string and convert it into a standard date format.

Let's consider the input strings below:

Company registered on 16 March 2003

or

Activity between 10 May 2006 an 10 July 2008 - no changes.

Now I would like a PHP function to apply over the strings and get the dates as YYYY-mm-dd

Example:

$date = DateExtract($sting1); // output: 2003-03-16

$date = DateExtract($sting2); // output: ['2006-05-10','2008-07-10']

Answer 1

For finding first two digit number Date Regexp would be - (?<![0-9])[0-9]{2}(?![0-9]) This can also be apply to four digit for Year also and for Month you can use hard-coded string search code.

 <?php

    $string = "Activity between 10 May 2006 an 10 July 2008 - no changes.";

    preg_match_all('/(\d{1,2}) (\w+) (\d{4})/', $string, $matches);
    print_r($matches);

    ?>

Output :-

    Array
(
    [0] => Array
        (
            [0] => 10 May 2006
            [1] => 10 July 2008
        )

    [1] => Array
        (
            [0] => 10
            [1] => 10
        )

    [2] => Array
        (
            [0] => May
            [1] => July
        )

    [3] => Array
        (
            [0] => 2006
            [1] => 2008
        )

)

Updates

For find Complete Date in string you can use this -

It works for short code for Month like Jan and complete name like January also.

Code -

<?php 
$string = "Activity between 10 May 2006 an 10 July 2008 - no changes.";

 preg_match_all('/(\b\d{1,2}\D{0,3})?\b(?:Jan(?:uary)?|Feb(?:ruary)?|Mar(?:ch)?|Apr(?:il)?|May|Jun(?:e)?|Jul(?:y)?|Aug(?:ust)?|Sep(?:tember)?|Oct(?:ober)?|(Nov|Dec)(?:ember)?)\D?(\d{1,2}\D?)?\D?((19[7-9]\d|20\d{2})|\d{2})/', $string, $complete);
    print_r($complete);
?>

Result -

Array
(
    [0] => Array
        (
            [0] => 10 May 2006
            [1] => 10 July 2008
        )

    [1] => Array
        (
            [0] => 10 
            [1] => 10 
        )

    [2] => Array
        (
            [0] => 
            [1] => 
        )

    [3] => Array
        (
            [0] => 20
            [1] => 20
        )

    [4] => Array
        (
            [0] => 06
            [1] => 08
        )

    [5] => Array
        (
            [0] => 
            [1] => 
        )

)

So you can fetch complete date form here and convert it into standard date format.

Demo Here

Rahul Dambare

Answer 2

Tricky. One approach might be to reason that dates always appear after certain grammatical words, as shown in your examples, e.g. "between", "on" etc. Using such words as a beginning anchor, we would then match until we find what we can reasonably assume to be the end of the date string. Here's what I hacked together:

//some strings
$strs = [
    "Company was in business between 14 March 2008 and 21 November 2012 inclusive",
    "I was born on 29 May 1980, 17:37 - it was a Thursday",
    "The big bang did not occur at 2pm, 14 Jun 1971, that's for sure."
];

//container to store possible date matches from strings
$possible_dates = array();

//prep months - long and short forms, to be used in matching
$date_prefix_words = array('between', 'on', 'at', 'during', 'and');
$months = array('January', 'February', 'March', 'April', 'May', 'June', 'July', 'August', 'September', 'October', 'November', 'December');
$months_short = array_map(function($month) { return substr($month, 0, 3); }, $months);

//iterate over and search strings - convert times like 2pm to 14:00, as, if they appear before the date, e.g. string 3, it doesn't get parsed
foreach($strs as $str) {
    $str = preg_replace_callback('/\b\d{1,2}[ap]m\b/', function($time) { return date('H:i', strtotime($time[0])); }, $str);
    preg_match_all('/(?<=\b'.implode('\b( |:)|\b', $date_prefix_words).'\b( |:))(\d|am|pm| |,|\'|:|'.implode('|', $months).'|'.implode('|', $months_short).')+/i', $str, $matches);
    if (count($matches)) $possible_dates = array_merge($possible_dates, $matches[0]);
}

//output before and after results
foreach($possible_dates as &$pd) {
    $pd = preg_replace('/, ?$/', '', $pd);
    echo '<p>Before: '.$pd.'<br />After: '.date('Y-m-d', strtotime($pd)).'</p>';
}

Clearly I'm making certain assumptions about your date formats, and you may need to tweak the REGEX, but it sort of works.

Answer 3

First of all You have to extract all part of date from the string separately.

RegEx Demo

---

First approach:

<?php

function standard_date_format($str) {
    preg_match_all('/(\d{1,2}) (\w+) (\d{4})/', $str, $matches);
    foreach ( $matches[1] as $day   ) { $days  [] = $day;   }
    foreach ( $matches[2] as $month ) { $months[] = $month; }
    foreach ( $matches[3] as $year  ) { $years [] = $year;  }

    $all_months = array('January', 'February', 'March', 'April', 'May', 'June', 'July', 'August', 'September', 'October', 'November', 'December');

    for ($i = sizeof ($days) - 1; $i >= 0; $i--) {
        $month     = array_search ($months[$i], $all_months) + 1;
        $month     = strlen ($month) < 2 ? '0'.$month : $month; 
        $results[] = $years[$i] . '-' . $month . '-' . $days[$i];
    }
    return  $results;
}

$str1 = "Company registered on 16 March 2003";
$str2 = "Activity between 10 May 2006 an 10 July 2008 - no changes.";

print_r(standard_date_format($str1)); // output: 2003-03-16
print_r(standard_date_format($str2)); // output: ['2006-05-10','2008-07-10']

PHP Demo

---

Second approach:

<?php

function standard_date_format($str) {
    preg_match_all('/(\d{1,2}) (\w+) (\d{4})/', $str, $matches);
    $dates  = array_map("strtotime", $matches[0]);
    $result = array_map(function($v) {return date("Y-m-d", $v); }, $dates);
    return $result;
}

$str1 = "Company registered on 16 March 2003";
$str2 = "Activity between 10 May 2006 an 10 July 2008 - no changes.";

print_r(standard_date_format($str1)); // output: 2003-03-16
print_r(standard_date_format($str2)); // output: ['2006-05-10','2008-07-10']

PHP Demo