多个关键字搜索，但有引号时完全匹配

Question

如果搜索vienna university，我的脚本会在某些MySQL 列中查找vienna 和university。

现在，我想通过以下方式对其进行修改：如果查询包含多个以引号开头和结尾的单词（例如"vienna university"），脚本不应分隔单词，而是搜索完全匹配的vienna university 在数据库中。

如果查询是Meyer "vienna university"，脚本应该在MySQL表中搜索Meyer和vienna university的存在。

您知道如何实现吗？

到目前为止我的代码：

<?php
$search_keyword = $_GET['query'];
//multiple keywords
$search_keyword = preg_split('/[\s]+/', $search_keyword);
$totalKeywords = count($search_keyword);
            
$sql = "SELECT * FROM `editors`
        WHERE CONCAT_WS(`journal`, `editor`, `affiliation`)
        LIKE ?";
        
//...for multiple keywords
for($i=1 ; $i < $totalKeywords; $i++){
    $sql .= " AND CONCAT_WS(`journal`, `editor`, `affiliation`)
              LIKE ?";
}
?>

Answer 1

这很容易用 PHP preg_* 函数完成；如果我们分步进行：

1. 匹配并从搜索查询中删除带引号的字符串
2. 用空格分割剩余的单词
3. 将两个结果合并在一起

代码

为方便起见，我将代码封装在一个函数中...

function splitWords($query){

    // Matches a single or double quote followed by any character  or more
    // times followed by the same quote as in group 1. i.e. if the first capture
    // a single quote then \1 == '
    $regex = '/(\'|")(.*?)\1/';

    // Find all matches
    // Matches for the second capture group will be returned in index [2]
    preg_match_all($regex, $query, $matches);

    // Remove quoted strings from the query
    // Split on spaces IF there's still words
    if( $query = trim(preg_replace($regex, '', $query)) )
        $arr = preg_split( "/\s+/", $query );

    // Return array of split words followed by quoted strings
    // $arr ?? [] ==> If $arr doesn't exist then $arr = []
    return array_merge($arr ?? [], $matches[2]);
}


print_r( splitWords("words 'in quotes' more letters 'and quotes' and \"other quotes\" ") );

print_r( splitWords('"Anna Huber"') );

print_r( splitWords('Anna') );

输出

Array
(
    [0] => words
    [1] => more
    [2] => letters
    [3] => and
    [4] => in quotes
    [5] => and quotes
    [6] => other quotes
)

Array
(
    [0] => Anna Huber
)

Array
(
    [0] => Anna
)

Answer 2

使用全文索引：

MATCH(col) AGAINST("vienna university") -- either word
MATCH(col) AGAINST("+vienna +university" IN BOOLEAN MODE) -- both words
MATCH(col) AGAINST('+"vienna university"') -- both words together

是的，您需要解析您的请求才能以这种方式重建它。是的，标点符号、停用词、最小字长等存在潜在问题。

速度很快。