MySql根据计数进行子选择和计数

Question

这是我的表结构模型。三张桌子。

----------------     ----------------------------     -------------------------------
|possibilities |     |realities                 |     |measurements                 |
|--------------|     |--------------------------|     |-----------------------------|
|pid| category |     |rid | pid | item | status |     |mid | rid | meas | date      |
|--------------|     |--------------------------|     |-----------------------------|
|1  | animal   |     |1   | 1   | dog  | 1 (yes)|     |1   | 1   | 3    | 2012-01-01|
|2  | vegetable|     |2   | 1   | fox  | 1      |     |2   | 3   | 2    | 2012-01-05|
|3  | mineral  |     |3   | 1   | cat  | 1      |     |3   | 1   | 13   | 2012-02-02|
----------------     |4   | 2   | apple| 2 (no) |     |4   | 3   | 24   | 2012-02-15|
                     |5   | 1   | mouse| 1      |     |5   | 2   | 5    | 2012-02-16|
                     |7   | 1   | bat  | 2      |     |6   | 6   | 4    | 2012-02-17|
                     ----------------------------     -------------------------------

我所追求的是一个结果，它将根据“可能性”表中特定条目的测量范围向我显示一系列计数，其中相关“现实”的状态为 1（意味着它当前正在被跟踪)，但唯一相关的测量是最近的测量。

这是我正在寻找的使用动物作为可能性的示例结果。

-----------------------
| 0-9 | 10-19 | 20-29 |
|---------------------|
| 2   | 1     | 1     |
-----------------------

因此，在此示例中，苹果行没有用于计数，因为它不是动物，蝙蝠也没有用于计数，因为它的状态设置为 no（意思是不测量），只有最近的测量值用于确定计数。

我目前在实际使用中有一个解决方法，但它没有遵循良好的数据库规范化。在我的现实表中，我有一个 current_meas 列，当进行新的测量并将其输入测量表时，该列会更新。然后我只需要使用前两个表，并且我有一个 SELECT 语句和一堆嵌入的 SUM 语句，例如使用 IF 值在 0-9 之间。它给了我想要的东西，但是我的应用已经发展到这种便利性在其他领域成为问题的地步。

那么，问题是，有没有更优雅的方式在一个语句中做到这一点？子选择？临时表？获取计数是该应用的核心。

这是一个基于 PHP5、MySQL5、JQuery 1.8 的 web 应用程序，以防我提供更多选择。提前致谢。我喜欢堆栈，并希望能提供尽可能多的帮助。

Answer 1

这是一种方法

创建一个临时表以获取最近的测量值

CREATE TEMPORARY TABLE RecentMeasurements
SELECT * FROM Measurements m
INNER JOIN (SELECT max(mid) max_id,date FROM Measurements GROUP BY DATE ORDER BY DATE ) x
ON x.max_id=m.mid

那你查询吗：

SELECT *, your counting logic
FROM Realities
WHERE status = 1 AND pid = 1
INNER JOIN RecentMeasurements

Answer 2

这是我根据建议的两个答案最终做的事情。

1. 首先我创建了一个临时表，它生成一个表 基于一种可能性（动物）的现实 状态为 1（是）。
2. 其次我创建了一个临时表，它生成一个表 从第一个临时表中的个人现实，并找到最 最近测量的每一项。
3. 从这第二个表中，我进行了一个选择，它给出了我的细分 在范围内计数。

当我仅使用一个临时表进行尝试时，每次查询都需要 5-10 秒。在我的实际使用中，我目前有 30 种可能性（一个脚本循环遍历每个可能性并生成这些临时表和选择），超过 1,000 个现实（在任何一天有 600 个处于活动状态，每月增加 100 个），以及超过 21,000 个测量值（每天添加 20-30 个）。那对我不起作用。因此，将其分解为较小的池以在 3-4 秒内将其从减少到运行整个报告。

这里是 MySQL 的东西，里面有我真实的表名和列名。

//Delete the temporary tables in advance
$delete_np_prod = 'DROP TABLE IF EXISTS np_infreppool';
mysql_query($delete_np_prod) or die ("Drop NP Prod Error " . mysql_error ());
$delete_np_max = 'DROP TABLE IF EXISTS np_maxbrixes';
mysql_query($delete_np_max) or die ("Drop NP Max Error " . mysql_error ());

//Make a temporary table to hold the totes of this product at North Plains that are active
$create_np_prod_pool_statement = 'CREATE TEMPORARY TABLE np_infreppool
SELECT inf_row_id FROM infusion WHERE formid = ' . $active_formids["formid"] . ' AND location = 1 AND status = 1';
mysql_query($create_np_prod_pool_statement) or die ("Prod Error " . mysql_error ());

//Make a temporary table to hold the tote with its most recent brix value attached to it.
$create_np_maxbrix_pool_statement = 'CREATE TEMPORARY TABLE np_maxbrixes
SELECT b.inf_row_id AS inf_row_id, b.brix AS brix from brix b, np_infreppool pool WHERE b.inf_row_id = pool.inf_row_id AND b.capture_date = (SELECT max(capture_date) FROM brix WHERE inf_row_id = pool.inf_row_id )';
mysql_query($create_np_maxbrix_pool_statement) or die ("Brix Error " . mysql_error ());

//Get the counts for slected form from NP
$get_report_np = "SELECT 
        SUM(IF(brix BETWEEN 0 AND 4,1,0)) as '0-4',
        SUM(IF(brix BETWEEN 5 AND 9,1,0)) as '5-9',
        SUM(IF(brix BETWEEN 10 AND 14,1,0)) as '10-14',
        SUM(IF(brix BETWEEN 15 AND 19,1,0)) as '15-19',
        SUM(IF(brix BETWEEN 20 AND 24,1,0)) as '20-24',
        SUM(IF(brix BETWEEN 25 AND 29,1,0)) as '25-29',
        SUM(IF(brix BETWEEN 30 AND 34,1,0)) as '30-34',
        SUM(IF(brix BETWEEN 35 AND 39,1,0)) as '35-39',
        SUM(IF(brix BETWEEN 40 AND 44,1,0)) as '40-44',
        SUM(IF(brix BETWEEN 45 AND 49,1,0)) as '45-49',
        SUM(IF(brix BETWEEN 50 AND 54,1,0)) as '50-54',
        SUM(IF(brix BETWEEN 55 AND 59,1,0)) as '54-49',
        SUM(IF(brix BETWEEN 60 AND 64,1,0)) as '60-64',
        SUM(IF(brix BETWEEN 65 AND 69,1,0)) as '65-69',
        SUM(IF(brix >=70, 1, 0)) as 'Over 70'
    FROM np_maxbrixes";
$do_get_report_np = mysql_query($get_report_np);
$got_report_np = mysql_fetch_array($do_get_report_np);

更新

我让它在单个 SELECT 语句中工作，而不使用临时表，它工作得更快。使用我上面的示例架构，它的外观如下。

    SELECT 
    SUM(IF(m.meas BETWEEN 0 AND 4,1,0)) as '0-4',
    SUM(IF(m.meas BETWEEN 5 AND 9,1,0)) as '5-9',
    SUM(IF(m.meas BETWEEN 10 AND 14,1,0)) as '10-14',
    SUM(IF(m.meas BETWEEN 15 AND 19,1,0)) as '15-19',
    SUM(IF(m.meas BETWEEN 20 AND 24,1,0)) as '20-24',
    SUM(IF(m.meas BETWEEN 25 AND 29,1,0)) as '25-29',
    SUM(IF(m.meas BETWEEN 30 AND 34,1,0)) as '30-34',
    SUM(IF(m.meas BETWEEN 35 AND 39,1,0)) as '35-39',
    SUM(IF(m.meas BETWEEN 40 AND 44,1,0)) as '40-44',
    SUM(IF(m.meas BETWEEN 45 AND 49,1,0)) as '45-49',
    SUM(IF(m.meas BETWEEN 50 AND 54,1,0)) as '50-54',
    SUM(IF(m.meas BETWEEN 55 AND 59,1,0)) as '54-49',
    SUM(IF(m.meas BETWEEN 60 AND 64,1,0)) as '60-64',
    SUM(IF(m.meas BETWEEN 65 AND 69,1,0)) as '65-69',
    SUM(IF(m.meas >=70, 1, 0)) as 'Over 70'
FROM measurement m, realities r 
WHERE r.status = 1 AND r.pid = " . $_GET['pid'] . " AND r.rid = m.rid AND m.date = (SELECT max(date) FROM measurements WHERE rid = r.rid)