Question

我想在一个查询中创建多个Z分数[（value-AVG）/ STD]并过滤它们。

到目前为止，我想出了这个;然而，在你看之前，有一些法语单词，“quartier”就像是一个区/自治市，“prix”是一个价格。

SELECT ((t1.prix-A1.HMEAN)*100/A1.STD_SAMPLE) AS Z,
q.myId AS q_myId
FROM immobilier_ad_blank AS t1 
LEFT JOIN Adresse AS a ON t1.adresse_id=a.id
LEFT JOIN Quartier AS q ON a.quartier_id=q.id
CROSS JOIN (
    SELECT AVG(t2.prix) AS HMEAN, 
    STD(t2.prix) AS STD_SAMPLE,
    q.myId AS quartier_myId
    FROM immobilier_ad_blank AS t2
    LEFT JOIN Adresse AS a ON t2.adresse_id=a.id
    LEFT JOIN Quartier AS q ON a.quartier_id=q.id
    GROUP BY quartier_myId
) AS A1 ON q.myId = A1.quartier_myId

它似乎有效，但我不确定我是否正确行事，我只添加了一个过滤条件，但我将添加最多6个条件，到目前为止，这个方法看起来非常麻烦。

最终看起来像这样

SELECT 
((t1.prix-A1.HMEAN)*100/A1.STD_SAMPLE) AS Z1,
((t1.prix-A2.HMEAN)*100/A1.STD_SAMPLE) AS Z2,
((t1.prix-AN.HMEAN)*100/A1.STD_SAMPLE) AS ZN,
All Relevant selects
FROM immobilier_ad_blank AS t1 
All Relevant LEFT Joins
CROSS JOIN (
    SELECT AVG(t2.prix) AS HMEAN, 
    STD(t2.prix) AS STD_SAMPLE,
    q.myId AS quartier_myId
    FROM immobilier_ad_blank AS t2
    LEFT JOIN Adresse AS a ON t2.adresse_id=a.id
    LEFT JOIN Quartier AS q ON a.quartier_id=q.id
    GROUP BY quartier_myId
) AS A1 ON q.myId = A1.quartier_myId
CROSS JOIN (
    SELECT AVG(t2.prix) AS HMEAN, 
    STD(t2.prix) AS STD_SAMPLE,
    q.myId AS quartier_myId
    FROM immobilier_ad_blank AS t3
    Other Joins
    Group By Other conditions
) AS A2 ON OtherConditions
CROSS JOIN (
    SELECT AVG(tN.prix) AS HMEAN, 
    STD(tN.prix) AS STD_SAMPLE,
    q.myId AS quartier_myId
    FROM immobilier_ad_blank AS tN
    Other Joins
    Group By Other conditions
) AS AN ON OtherConditions

这是在mysql中，根据手工交叉连接=内连接=连接。但是根据我的理解，交叉连接是表X表，但在我的情况下，因为它返回聚合，我想它不是t ^ 2而是t *（聚合数）对吗？

我正确地走在正确的轨道上？优化明智？我把过滤器放在了正确的位置吗？

我基本上想要选择所有行，找到它们的连接值的平均值函数，计算它们的Z分数，并在不到5秒的时间内将它们加起来并且有大约1M的数据。

编辑1：

我会简化我的问题，

SELECT 
((t1.prix-A1.HMEAN)*100/A1.STD_SAMPLE) AS Z,
((t1.prix-A2.HMEAN)*100/A2.STD_SAMPLE) AS Z2,
q.myId AS q_myId,
c.myId AS c_myId,
s.myId as size_myId
FROM immobilier_ad_blank AS t1 
LEFT JOIN Size AS s ON t1.size_id=s.id
LEFT JOIN Adresse AS a ON t1.adresse_id=a.id
LEFT JOIN City AS c ON a.city_id=c.id
LEFT JOIN Quartier AS q ON a.quartier_id=q.id
CROSS JOIN (
    SELECT AVG(t2.prix) AS HMEAN, 
    STD(t2.prix) AS STD_SAMPLE,
    q.myId AS quartier_myId
    s.myId as size_myId
    FROM immobilier_ad_blank AS t2
    LEFT JOIN Size AS s ON t2.size_id=s.id
    LEFT JOIN Adresse AS a ON t2.adresse_id=a.id
    LEFT JOIN Quartier AS q ON a.quartier_id=q.id
    GROUP BY quartier_myId, size_myId
) AS A1 ON q.myId = A1.quartier_myId
AS A2 on s.myId=A2.size_myId #<--------- Is this line possible ?

编辑3：

我最终使用临时表并复制它们，因为我的情况下的视图比临时表慢，即使基础数据被正确编入索引。

CREATE TEMPORARY TABLE IF NOT EXISTS A1 AS (
SELECT AVG(t2.prix) AS HMEAN, 
    STD(t2.prix) AS STD_SAMPLE,
    s.myId AS size_myId,
    q.myId AS quartier_myId
    FROM immobilier_ad_blank AS t2
    LEFT JOIN Size AS s ON t2.size_id=s.id
    LEFT JOIN Adresse AS a ON t2.adresse_id=a.id
    LEFT JOIN Quartier AS q ON a.quartier_id=q.id
    GROUP BY quartier_myId,size_myId);

CREATE TEMPORARY TABLE A2 LIKE A1; 
INSERT A2 SELECT * FROM A1;

SELECT 
((c.prix-A1.HMEAN)*100/A1.STD_SAMPLE) AS Z1,
((c.prix-A2.HMEAN)*100/A2.STD_SAMPLE) AS Z2,
q.myId AS quartier_myId,
s.myId AS size_myId
FROM immobilier_ad_blank AS c
LEFT JOIN Size AS s ON c.size_id=s.id
LEFT JOIN Adresse AS ad ON c.adresse_id=ad.id
LEFT JOIN Quartier AS q ON ad.quartier_id=q.id
JOIN A1 on A1.quartier_myId = q.myId
JOIN A2 AS A2 on A2.size_myId = s.myId

基本上，加入相同的虚拟表（子查询）=== Views ||临时表说实话，我最终可能会创建一个永久表并每天更新一次....

Answer 1

CROSS JOIN ... ON condition使交叉联接成为普通的INNER JOIN，或者为简洁起见，只需JOIN。所以不要担心组合爆炸。正确编写的ON条件会阻止它。简而言之，忘记CROSS JOIN。只需写下JOIN。

您可以将JOIN写成类似这样的内容

 FROM table_a AS a
 JOIN table_b AS b ON a.id = b.id

或者，您可以将table_a或table_b或两者表示为子查询。例如，您可以编写此查询。

 FROM table_A AS a
 JOIN (
         SELECT MAX(id) id, district
           FROM table_b
          GROUP BY district
      ) AS b on a.id = b.id

换句话说，您可以互换地使用物理表（table_b）或虚拟表（子查询）。

在您的情况下，您的子查询是一个四列表：（HMEAN，STD_SAMPLE，quartier_myId和size_myId）。

如何将该表加入查询中的其余表？

你有代码

  JOIN (subquery) AS A1 ON q.myId = A1.quartier_myId

但是，要使用子查询的第四列（size_myId），还需要将它用于JOIN。为此，请在AND子句中加上ON。做这样的事情：

  JOIN (subquery) AS A1   ON q.myId = A1.quartier_myId 
                         AND s.myId = A.size_myId

如果s.myId的含义是正确的，那应该会给你一个有用的结果。

复合ON子句非常有用。你可以这样做：

  LEFT JOIN (subquery) AS A1   ON q.myId = A1.quartier_myId 
                              AND s.myId = A.size_myId
                              AND A.HMEAN > 7.5

过滤结果。

多个SubSelect内部连接Mysql

1 个答案: