Php 优化MySQL搜索查询

我只是尝试对我的网站进行一些小搜索，以找到用户条目并以逻辑方式对它们进行排序，我尝试改进结果，但结果是查询字符串太大，需要14秒才能在我的服务器上执行。因此，我现在想知道如何在不改变结果的情况下优化查询

以下是我当前的查询：

$query = "SELECT * FROM(
    SELECT 1 as m_rank,title,rank, content, url, keywords FROM websites
        WHERE title LIKE '".$search."'
            union all
    SELECT 2 as m_rank,title,rank, content, url, keywords FROM websites
        WHERE title LIKE '".$search."%'
        OR title LIKE '%".$search."'
            union all
    SELECT 3 as m_rank,title,rank, content, url, keywords FROM websites
        WHERE title LIKE '%".$search."%'
            union all
    SELECT 4 as m_rank,title,rank, content, url, keywords FROM websites
        WHERE keywords LIKE '".$search."'
            union all
    SELECT 5 as m_rank,title,rank, content, url, keywords FROM websites
        WHERE keywords LIKE '%".$search."'
        OR keywords LIKE '".$search."%'
            union all
    SELECT 6 as m_rank,title,rank, content, url, keywords FROM websites
        WHERE keywords LIKE '%".$search."%'
            union all
    SELECT 7 as m_rank,title,rank, content, url, keywords FROM websites
        WHERE content LIKE '".$search."'
            union all
    SELECT 8 as m_rank,title,rank, content, url, keywords FROM websites
        WHERE content LIKE '%".$search."'
        OR content LIKE '".$search."%'
            union all
    SELECT 9 as m_rank,title,rank, content, url, keywords FROM websites
        WHERE content LIKE '%".$search."%'
    ) a
    ORDER BY m_rank, rank DESC LIMIT ".$page.", ".$max;

---

是否也可以使用PHP进行预处理来剪切查询？

您可以通过一个查询来完成此操作。将不同的搜索条件放入

案例中

，以分配

m_秩

。比如：

SELECT title, rank, content, url, keywords,
       (case when title LIKE '".$search."' then 1
             when title LIKE '".$search."%' OR title LIKE '%".$search."' then 2
             when title LIKE '%".$search."%' then 3
             . . .
             else NULL
        end) as m_rank
FROM websites
HAVING m_rank is not null
ORDER BY m_rank, rank DESC
LIMIT ".$page.", ".$max;

---

很好-使用

CASE

（）比我以前写过的一些

IF

庞然大物要整洁得多：）其他可能的替代方法：

FULLTEXT

搜索（）或实现类似ApacheLucene（）的东西，而不是RYO解决方案。