Question

这是我当前的查询：

SELECT q.type,
      q.author_id author,
      q.deleted,
      q.deleter_id,
      COUNT(DISTINCT a.id) answers_num,
      COUNT(DISTINCT v.id) votes_num,
FROM qanda q
LEFT  JOIN qanda a ON q.id = a.related
LEFT  JOIN votes v on q.id = v.post_id
WHERE q.id = ?

效果很好。我只需要在EXISTS语句中再添加一件事（由SELECT返回一个布尔值）。这个：

EXISTS (SELECT a.id, sum(vot.value) total_votes
        FROM qanda q
        LEFT JOIN qanda a on q.id = a.related
        LEFT JOIN votes vot ON a.id = vot.post_id
        WHERE q.type = 0         -- this is a question
          AND a.related = 1315   -- getting answers
        GROUP BY a.id
        HAVING total_votes
        ORDER BY total_votes DESC
        LIMIT1
) as HasAUpvotedAnswer

好的，它也很好用。但我担心性能。如您所见，以上两个查询中的JOIN（都是最终合并的）是相似的。如何使它们更优化？

换句话说，如何在最佳情况下结合以上两个查询？

Answer 1

查看您的代码可能是y可以使用条件查询求和的单个查询

SELECT q.type,
      q.author_id author,
      q.deleted,
      q.deleter_id,
      COUNT(DISTINCT a.id) answers_num,
      COUNT(DISTINCT v.id) votes_num,
      IF (SUM(
        CASE WHEN q.type  = 0 AND a.related = 1315
          THEN v.value  else  0 END) > 0,1,0) 
      HasAUpvotedAnswer
FROM qanda q
LEFT  JOIN qanda a ON q.id = a.related
LEFT  JOIN votes v on q.id = v.post_id
WHERE q.id = ?

如何减少子查询中的JOIN？

1 个答案: