meta_query search names when they have middle initials - PHP, WordPress

Question

I have to build a search form where I search members of the company by their Full Name (First Name + Middle Initial + Last Name). All the names are in this form: John B. Doe.

The below code is working on these cases: John, John B., Doe, B., John B. Doe, but not working the query search is: John Doe.

 if (isset($_POST['search']) && $_POST['search'] != -1) {
    $args['meta_query']['name_search']['key'] = "_full_name";
    $args['meta_query']['name_search']['compare'] = "LIKE";
    $args['meta_query']['name_search']['value'] = $_POST['search'];
  }

How should I improve the query in order to work with both: the name and surname (John Doe) and the name, surname plus the middle initial (John B. Doe)?

Answer 1

First off, if only WordPress doesn't escape the % characters in the value (i.e. the search keyword), you could've simply replaced spaces in the search keyword with %, hence you'd get a clause like meta_value LIKE '%John%Doe%' which would match John B. Doe.

So because the % in the search keyword is being escaped (which is a good thing, BTW), then you can instead use REGEXP (regular expression search) and then replace the spaces with a .* which is equivalent to the % in a LIKE clause.

Working Example

Replace this in your code:

$args['meta_query']['name_search']['compare'] = "LIKE";
$args['meta_query']['name_search']['value'] = $_POST['search'];

with this:

// Build the regular expression pattern.
$list = preg_split( '/ +/', trim( $_POST['search'] ) );
$regex = implode( '.*', array_map( 'preg_quote', $list ) );

// Then set 'compare' to REGEXP and 'value' to the above pattern.
$args['meta_query']['name_search']['compare'] = 'REGEXP';
$args['meta_query']['name_search']['value'] = $regex;

Tried & tested working in WordPress 5.7.2, but do take note of the "not multibyte safe" warning in the MySQL reference manual.

Alternate Solution (equivalent to the one above)

If you want to use LIKE and not REGEXP, but don't want to end up with potentially lots of meta queries for the same key, then you can:

Use three search fields, namely first name, last name, and middle initial, and three meta clauses, one for each search field, but all having key set to _full_name. E.g.

/*
 * Assuming your form contains these:
    <input name="first_name">
    <input name="last_name">
    <input name="middle_initial">
 */

$name_search = array(
    'relation' => 'AND',
);

// * okay, this still uses 3 clauses, but it always would be just 3 clauses
foreach ( array( 'first_name', 'last_name', 'middle_initial' ) as $name ) {
    if ( ! empty( $_POST[ $name ] ) ) {
        $name_search[ $name ] = array(
            'key'     => '_full_name',
            'value'   => sanitize_text_field( $_POST[ $name ] ),
            'compare' => 'LIKE',
        );
    }
}

$args['meta_query']['name_search'] = $name_search;

Answer 2

I can suggest the following solution:

$search = $_POST['search'];
$words  = explode( $search, ' ' );
$words  = array_filter(
    $words,
    function ( $word ) {
        return ! empty( trim( $word ) );
    }
);

foreach ( $words as $index => $word ) {
    $query_name          = "name_search_$index";
    $args['meta_query'][ $query_name ] = [
        'key'     => '_full_name',
        'value'   => $word,
        'compare' => 'LIKE',
    ];
}

Code explanation:

1. I split search phrases into words. Words are any sequence of symbols between spaces
2. I remove spaces. If your search string is

   John Doe

   or

   John  Doe

   or

      John        Doe   

   The search result should be the same.
3. I build multiple queries. Every query corresponds to a single word. If you search John Doe the query builds by 2 queries. Queries count as much as the number of words.

The problems with this solution:

1. Too many words are too many queries. Too many LIKE-containing queries are too slow.
2. You will get the same result for searching for

   John Doe

   and

   Doe John

   . The order of the words makes no difference to the result.

---

I recommend using this solution only as a temporary solution. Good solution - is using Full-Text Search plugins or write Full-Text Search solution yourself. More details about FullText search you could find on MySQL official page: https://dev.mysql.com/doc/refman/8.0/en/fulltext-search.html