Php 如何在全文MySQL搜索中同时匹配HtmlLentified关键字和非HtmlLentified关键字

我学会了在MySQL中尽可能原始地保存数据。下面是我如何在MySQL数据库中存储内容的示例：

title (VARCHAR, 255) => Références
content (TEXT) => <p>A paragraph about r&eacute;f&eacute;rences...</p>

您也可以在

内容中保留
“référence”
不替换，因为它仍然是正确的HTML

（只要HTML头指定的编码与数据库中的编码相同，假定为UTF-8：

<meta http-equiv="Content-Type" content="text/html;charset=utf-8" >



)对于全文搜索，我相信您可以在表达式中使用
或
运算符。有关更多详细信息，请参阅。您可以在一个查询中使用它

在您的代码中，前面提到的内容与此类似：

$keyword = utf8_decode(mysql_real_escape_string($_POST['keyword'])); 
$keyword = '(' . $keyword . '*) OR (' . htmlentities($keyword) . '*)';

（有点脏，但我想你可以把它清理干净；）

我不知道在您的情况下，
$keyword
变量是否可以包含多个单词。所以我加了括号来澄清这一点

另一方面：我不确定布尔全文搜索是否会看到类似
ré；fé；法语
作为一个单词。但我想值得一试；）

编辑：我刚刚发现前面提到的问题可以通过在关键字周围使用双qoutes来解决，例如：
“ré；fé；rences”
。也就是说，如果有偏离路线的问题；）
 作为一名在法国网站工作的法国开发人员，我也遇到了同样的问题，并找到了一种解决方法，同时仍然使用我认为您会发现有用的强大匹配功能。
你可以原谅可能的英语错误，我说我是法国人

我遇到的问题是，“HTMLlentified”这个词没有被识别为一个单词。例如，一旦在数据库中出现“HTMLTENTIFECT”，并且与函数匹配时将其视为（三个不同的单词）。

但正如上面所说的，引号之间的“htmlentified”单词（“）本身被视为一个表达式，并且被视为（一个表达式）

因此，解决方案似乎是，我需要保持表达式在引号之间的原样，并使用正则表达式在引号之间放置“htmlensified”单词

下面是我为解决这个问题而开发的代码，直到有人修复MySQL中的错误

    //At start, my variable is "htmlentified", so I need to replace <<&quot;>> with real double quotes <<">>
$Search_Text=str_replace("&quot;","\"",$Search_Text);

//I used the "ENT_COMPAT" option for html_entities, so I need to escape the single quotes
$Search_Text=str_replace("'","\\'",$Search_Text);

//I search if there are text between double quotes using "preg_match_all" with "PREG_OFFSET_CAPTURE" option to have starting offset for each of them
//Note that at the start of the regex I keep the eventual "+" or "-" or "(" or ")" just before or after quotes
if(preg_match_all('#([-|\+|(]{0,})(\"(.+)\")([)]{0,})#isU',$Search_Text,$Quoted_Blocks,PREG_OFFSET_CAPTURE))
{
    //I Initialise an array that will contain each distinct words and quoted expressions from the search text
    $List_Search_Terms=Array();

    //If the offset of the first quoted expression is not 0, it means that there is some words before that I will put in the array
    if($Quoted_Blocks[0][0][1]>0)
    {
        $Starting_Offset=0;
        $Length=$Quoted_Blocks[0][0][1];

        foreach(explode(" ",trim(substr($Search_Text,$Starting_Offset,$Length))) as $Index=>$Valeur)
        {
            array_push($List_Search_Terms,$Valeur);
        }
    }

    //Then I treat all quoted expressions and words between them that I will put in the array
    for($Index_Bloc=0;$Index_Bloc<count($Quoted_Blocks[0]);$Index_Bloc++)
    {
        //I put the quoted expression unmodified in my array
        array_push($List_Search_Terms,$Quoted_Blocks[0][$Index_Bloc][0]);

        //I verify if there's a following quoted expression for I can treat the words in between
        if(isset($Quoted_Blocks[0][$Index_Bloc+1]))
        {
            $Starting_Offset=$Quoted_Blocks[0][$Index_Bloc][1]+strlen($Quoted_Blocks[0][$Index_Bloc][0]);
            $Length=$Quoted_Blocks[0][$Index_Bloc+1][1]-$Starting_Offset;

            //I put the words in between in the array
            foreach(explode(" ",trim(substr($Search_Text,$Starting_Offset,$Length))) as $Index=>$Valeur)
            {
                array_push($List_Search_Terms,$Valeur);
            }
        }
    }

    //After treating all quoted expressions, I put in the array the words that can remain after the last quoted expression
    $Starting_Offset=$Quoted_Blocks[0][count($Quoted_Blocks[0])-1][1]+strlen($Quoted_Blocks[0][count($Quoted_Blocks[0])-1][0]);

    if($Starting_Offset<strlen($Search_Text))
    {
        foreach(explode(" ",trim(substr($Search_Text,$Starting_Offset))) as $Index=>$Valeur)
        {
            array_push($List_Search_Terms,$Valeur);
        }
    }
}
else
{
    //If there's no quoted expression in the search, I put all words in the array
    $List_Search_Terms=explode(" ",$Search_Text);
}

//Once I have all words and quoted expressions in the array, I can run through it to put "htmlentified" words (not expressions !) between double quotes
foreach($List_Search_Terms as $Index=>$Value)
{
    //I control that the line of the array doesn't contain double-quotes (which I won't touch) and if there are "&" and ";" characters that means an "htmlentified" word
    if(!preg_match('#"#',$Value) && preg_match('#&(.+)\;#isU',$Value))
    {
        //if both conditions match, I put quotes at the beginning and end
        //Note that as abovre, the regex will keep the eventual "+" or "-" or "(" or ")" just before or after the word
        $List_Search_Terms[$Index]=preg_replace('#^([-|\+|(]{0,})(.+)([)]{0,})$#is',"$1\"$2\"$3",$Value);
    }
}

//After that treatment, the array is converted in one single text that can be included in the mysql AGAINST condition
$Final_Search_Text=implode(" ",$List_Search_Terms);

//开始时，我的变量是“htmlentified”，所以我需要替换>
$Search\u Text=str\u replace（“”，“\”，$Search\u Text）；
//我对html_实体使用了“ENT_COMPAT”选项，因此需要对单引号进行转义
$Search\u Text=str\u replace（“'”、“\\'”、$Search\u Text）；
//我使用带有“preg\u OFFSET\u CAPTURE”选项的“preg\u match\u all”搜索双引号之间是否有文本，以使每个双引号都有起始偏移量
//请注意，在正则表达式的开头，我保留了引号之前或之后的最终“+”或“-”或“（“或”）”
如果（preg|u match|u all（'.|（[-|\+|（]{0，}）（\“（.+）\”）（[）]{0，}）\#isU'，$Search_Text，$Quoted|u块，preg|u偏移量_捕获））
{
//我初始化了一个数组，该数组将包含搜索文本中每个不同的单词和引用的表达式
$List_Search_Terms=Array（）；
//如果第一个带引号的表达式的偏移量不是0，这意味着在我将放入数组之前有一些单词
如果（$Quoted_Blocks[0][0][1]>0）
{
$Starting_Offset=0；
$Length=$Quoted_块[0][0][1]；
foreach（explode（“），trim（substr（$Search\u Text，$start\u Offset，$Length）），作为$Index=>$Valeur）
{
数组推送（$List\u Search\u Terms，$Valeur）；
}
}
//然后我处理所有引用的表达式和它们之间的单词，我将把它们放入数组中
对于（$Index\u Bloc=0；$Index\u Bloc$Value）
{
//我控制数组的行不包含双引号（我不想碰它），如果有“&”和“；”字符，则表示“htmlentified”字
if（！preg#u match（'#“#，$Value）&&preg_match（'#和（.+）\#isU'，$Value））
{
//如果两个条件都匹配，我会在开头和结尾加引号
//注意，作为abovre，regex将保留单词前后的最终“+”或“-”或“（“或”）”
$List_Search_Terms[$Index]=preg_replace（“#^（[-|\+|（]{0，}）（.+）（[）]{0，}）$#is'，“$1\”$2\“$3”，$Value）；
}
}
//在处理之后，数组被转换成一个文本，该文本可以包含在mysql条件中
$Final_Search_Text=内爆（“，$List_Search_Terms”）；

人们可能会对它进行一些优化，但事实上，它确实起到了作用！
谢谢。这意味着我不再使用
htmlentities
在页面上输出任何内容了？这不是有点不安全吗？只是出于好奇：我有了这个
AddDefaultCharset utf-8
和
AddCharset utf-8.html.css.js.xml.json.rss
在我的.htaccess中。另外，在我的HTML页面顶部我有
。这些东西之间有什么区别，如果有的话？在输出标题时，你仍然会使用htmlentities。在生成
内容时，我会htmlentity它，然后重新替换所需的字符（“白名单方法”：é；->。对于你的第二个问题：Apache也知道它是UTF-8，这很好。其含义与上面的元标记相同（我以前没有看到它的变体…），但对于跨浏览器的兼容性，你不能经常声明它…此外，PHP（默认字符集）、mysql驱动程序（mysql\U集\U字符集（））以及mysql（在数据库、表和列级别）可以/应该设置为utf-8。这样，您可以避免写入utf8_decode（），因为它的存储方式与输出方式相同。感谢您对utf-8一致性的见解。就我的数据库而言，它设置为utf-8。我在PHP中的默认字符集设置为“无值”。我可能可以使用.htaccess来更改此设置。我不能100%确定您关于htmlentity ing
content
的评论。这不会弄乱HTML标记，如
pA段落…
htmlspecialchars在这种情况下是否有用？非常感谢，但这似乎会给我太多的误报。正如您所指出的布莱
    //At start, my variable is "htmlentified", so I need to replace <<&quot;>> with real double quotes <<">>
$Search_Text=str_replace("&quot;","\"",$Search_Text);

//I used the "ENT_COMPAT" option for html_entities, so I need to escape the single quotes
$Search_Text=str_replace("'","\\'",$Search_Text);

//I search if there are text between double quotes using "preg_match_all" with "PREG_OFFSET_CAPTURE" option to have starting offset for each of them
//Note that at the start of the regex I keep the eventual "+" or "-" or "(" or ")" just before or after quotes
if(preg_match_all('#([-|\+|(]{0,})(\"(.+)\")([)]{0,})#isU',$Search_Text,$Quoted_Blocks,PREG_OFFSET_CAPTURE))
{
    //I Initialise an array that will contain each distinct words and quoted expressions from the search text
    $List_Search_Terms=Array();

    //If the offset of the first quoted expression is not 0, it means that there is some words before that I will put in the array
    if($Quoted_Blocks[0][0][1]>0)
    {
        $Starting_Offset=0;
        $Length=$Quoted_Blocks[0][0][1];

        foreach(explode(" ",trim(substr($Search_Text,$Starting_Offset,$Length))) as $Index=>$Valeur)
        {
            array_push($List_Search_Terms,$Valeur);
        }
    }

    //Then I treat all quoted expressions and words between them that I will put in the array
    for($Index_Bloc=0;$Index_Bloc<count($Quoted_Blocks[0]);$Index_Bloc++)
    {
        //I put the quoted expression unmodified in my array
        array_push($List_Search_Terms,$Quoted_Blocks[0][$Index_Bloc][0]);

        //I verify if there's a following quoted expression for I can treat the words in between
        if(isset($Quoted_Blocks[0][$Index_Bloc+1]))
        {
            $Starting_Offset=$Quoted_Blocks[0][$Index_Bloc][1]+strlen($Quoted_Blocks[0][$Index_Bloc][0]);
            $Length=$Quoted_Blocks[0][$Index_Bloc+1][1]-$Starting_Offset;

            //I put the words in between in the array
            foreach(explode(" ",trim(substr($Search_Text,$Starting_Offset,$Length))) as $Index=>$Valeur)
            {
                array_push($List_Search_Terms,$Valeur);
            }
        }
    }

    //After treating all quoted expressions, I put in the array the words that can remain after the last quoted expression
    $Starting_Offset=$Quoted_Blocks[0][count($Quoted_Blocks[0])-1][1]+strlen($Quoted_Blocks[0][count($Quoted_Blocks[0])-1][0]);

    if($Starting_Offset<strlen($Search_Text))
    {
        foreach(explode(" ",trim(substr($Search_Text,$Starting_Offset))) as $Index=>$Valeur)
        {
            array_push($List_Search_Terms,$Valeur);
        }
    }
}
else
{
    //If there's no quoted expression in the search, I put all words in the array
    $List_Search_Terms=explode(" ",$Search_Text);
}

//Once I have all words and quoted expressions in the array, I can run through it to put "htmlentified" words (not expressions !) between double quotes
foreach($List_Search_Terms as $Index=>$Value)
{
    //I control that the line of the array doesn't contain double-quotes (which I won't touch) and if there are "&" and ";" characters that means an "htmlentified" word
    if(!preg_match('#"#',$Value) && preg_match('#&(.+)\;#isU',$Value))
    {
        //if both conditions match, I put quotes at the beginning and end
        //Note that as abovre, the regex will keep the eventual "+" or "-" or "(" or ")" just before or after the word
        $List_Search_Terms[$Index]=preg_replace('#^([-|\+|(]{0,})(.+)([)]{0,})$#is',"$1\"$2\"$3",$Value);
    }
}

//After that treatment, the array is converted in one single text that can be included in the mysql AGAINST condition
$Final_Search_Text=implode(" ",$List_Search_Terms);