php - 快速加入大数据表与键表

Question

这是我的第二个问题，所以请耐心等待:)

我有一个 MySQL 数据库，其中有一个大表 ( dati)，结构如下:

CREATE TABLE IF NOT EXISTS `dati` (
  `i1` varchar(200) COLLATE utf8_swedish_ci DEFAULT NULL,
  `i2` varchar(200) COLLATE utf8_swedish_ci DEFAULT NULL,
  `i3` varchar(200) COLLATE utf8_swedish_ci DEFAULT NULL,
  `i4` varchar(200) COLLATE utf8_swedish_ci DEFAULT NULL,
  `i5` varchar(200) COLLATE utf8_swedish_ci DEFAULT NULL,
  `i6` varchar(200) COLLATE utf8_swedish_ci DEFAULT NULL,
  `i7` varchar(200) COLLATE utf8_swedish_ci DEFAULT NULL,
  `i8` varchar(200) COLLATE utf8_swedish_ci DEFAULT NULL,
  `i9` varchar(200) COLLATE utf8_swedish_ci DEFAULT NULL,
  `i10` varchar(200) COLLATE utf8_swedish_ci DEFAULT NULL,
  `totale` double DEFAULT NULL,
  `valore` double DEFAULT NULL,
  KEY `i1` (`i1`(20),`i2`(20),`i3`(20),`i4`(20)),
  KEY `i1_2` (`i1`),
  KEY `i2` (`i2`),
  KEY `i3` (`i3`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_swedish_ci;

和一些小 table i1 - i4其中字段中的值 valore链接到 data 中的值. i<n>具有这种结构:

CREATE TABLE IF NOT EXISTS `i1` (
  `livello` varchar(100) COLLATE utf8_swedish_ci DEFAULT NULL,
  `valore` varchar(100) COLLATE utf8_swedish_ci DEFAULT NULL,
  `ordine` double DEFAULT NULL,
  `colore` double DEFAULT NULL,
  `mostrare` double DEFAULT NULL,
  KEY `livello` (`livello`),
  KEY `valore` (`valore`),
  KEY `ordine` (`ordine`),
  KEY `mostrare` (`mostrare`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_swedish_ci;

CREATE TABLE IF NOT EXISTS `i2` (
  `livello` varchar(100) COLLATE utf8_swedish_ci DEFAULT NULL,
  `valore` varchar(100) COLLATE utf8_swedish_ci DEFAULT NULL,
  `ordine` double DEFAULT NULL,
  `colore` double DEFAULT NULL,
  `mostrare` double DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_swedish_ci;

CREATE TABLE IF NOT EXISTS `i3` (
  `livello` varchar(100) COLLATE utf8_swedish_ci DEFAULT NULL,
  `valore` varchar(100) COLLATE utf8_swedish_ci DEFAULT NULL,
  `ordine` double DEFAULT NULL,
  `colore` double DEFAULT NULL,
  `mostrare` double DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_swedish_ci;

CREATE TABLE IF NOT EXISTS `i4` (
  `livello` varchar(100) COLLATE utf8_swedish_ci DEFAULT NULL,
  `valore` varchar(100) COLLATE utf8_swedish_ci DEFAULT NULL,
  `ordine` double DEFAULT NULL,
  `colore` double DEFAULT NULL,
  `mostrare` double DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_swedish_ci;

对于每个 i1-14，我都有一个由用户做出的选择，并与用户对其允许值的权限合并，PHP 生成一个查询，格式如下:

SELECT * FROM `dati`
WHERE (`i1` = 'TARGET|TOTALE')
AND (`i2` = 'BANCHE|ADV AWARENESS SPONTANEA OFFLINE')
AND (`i3` = 'BRAND BANCARI|(NET) BANCOPOSTA')
AND (`i4` = 'ANNO|2014' OR `i4` = 'ANNO|2013')

我的问题是我需要所有键组合的值，即使是那些在主数据表中没有条目的组合。

有人建议我使用 i1-i4 表的 LEFT JOIN 将获取所有条目，但我尝试了很多方法，但总是遇到错误或非常缓慢或永无止境的查询。如何加入dati与 i1 - i4以快速的方式，以便所有(选定的)组合键都出现(即使数据中没有条目)。

所以我需要加入 dati.i1=i1.valore、dati.i2=i2.valore 等。

感谢您的帮助。 (P.S 我不会再关注这个几个小时了，现在是晚上，稍后会在家回来)

编辑

现在在家，我将尝试发布一些可能有帮助的示例。

这是一个客户端脚本的查询，适用于这些表的名称，我不知道他们为什么这样做，它似乎以某种方式工作，虽然它对我来说看起来过于复杂，必须生成它PHP 动态:

select k.i4 , k.i3, k.i1, k.i2, dati.valore, dati.totale as totale, s.base
from (( select * from (
select valore as i4, ordine as ordine_i4 from i4 where valore in ('ANNO|2013','ANNO|2014')) p,
(select valore as i3, ordine as ordine_i3 from i3 where valore in ('BRAND BANCARI|(NET) BANCOPOSTA','BRAND BANCARI|(NET) CONTO ARANCIO/ING DIRECT','BRAND BANCARI|(NET) GRUPPO UNICREDIT','BRAND BANCARI|BANCA MEDIOLANUM')) b,
(select valore as i1, ordine as ordine_i1 from i1 where valore in ('TARGET|INTERNAUTI','TARGET|INTERNAUTI SOCIAL')) t,
(select valore as i2, ordine as ordine_i2 from i2 where valore in ('BANCHE|ADV AWARENESS ONLINE TOTALE','BANCHE|BRAND AWARENESS TOM','BANCHE|BRAND AWARENESS TOTALE','BANCHE|BRAND AWARENESS TOTALE SPONTANEA','BANCHE|NOTIZIABILITA\' OFFLINE','BANCHE|NOTIZIABILITA\' OFFLINE + ONLINE','BANCHE|NOTIZIABILITA\' ONLINE')) a
 ) k left JOIN (
 select * from dati where i4 in ('ANNO|2013','ANNO|2014') And i3 in ('BRAND BANCARI|(NET) BANCOPOSTA','BRAND BANCARI|(NET) CONTO ARANCIO/ING DIRECT','BRAND BANCARI|(NET) GRUPPO UNICREDIT','BRAND BANCARI|BANCA MEDIOLANUM') And i1 in ('TARGET|INTERNAUTI','TARGET|INTERNAUTI SOCIAL') And i2 in ('BANCHE|ADV AWARENESS ONLINE TOTALE','BANCHE|BRAND AWARENESS TOM','BANCHE|BRAND AWARENESS TOTALE','BANCHE|BRAND AWARENESS TOTALE SPONTANEA','BANCHE|NOTIZIABILITA\' OFFLINE','BANCHE|NOTIZIABILITA\' OFFLINE + ONLINE','BANCHE|NOTIZIABILITA\' ONLINE')) as dati
on k.i4=dati.i4 and k.i3=dati.i3 and k.i1=dati.i1 and k.i2=dati.i2 ) left JOIN (select i4, i1, i2, valore as base from dati where i1='TOTALE' ) s on k.i4=s.i4 and k.i1=s.i1 and k.i2=s.i2 ORDER BY k.ordine_i3, k.ordine_i4, k.ordine_i2, k.ordine_i1

还有我的一些NOT WORKING 试验。这在某处有明显的错误，但我似乎无法修复它

((SELECT * FROM `dati`
    WHERE (`i1` = 'TARGET|TOTALE')
    AND (`i2` = 'BANCHE|ADV AWARENESS SPONTANEA OFFLINE')
    AND (`i3` = 'BRAND BANCARI|(NET) BANCOPOSTA')
    AND (`i4` = 'ANNO|2014')
) aaa )

LEFT JOIN (

    SELECT * FROM (

        (SELECT valore AS `vi1`
        FROM `i1`
        ) `ti1`,

        (SELECT valore AS `vi2`
        FROM `i2`
        ) `ti2`,

        (SELECT valore AS `vi3`
        FROM `i3`
        ) `ti3`,

        (SELECT valore AS `vi4`
        FROM `i4`
        ) `ti4`

    ) bbb
)  ON `ti1`.`vi1` = `dati`.`i1` AND `ti2`.`vi2` = `dati`.`i2`  AND `ti3`.`vi3` = `dati`.`i3`  AND `ti4`.`vi4` = `dati`.`i4`

这个看起来很酷，但只是永远锁定 MySQL(!!!):

( select * from


( select * from
    (select valore as `vi1`,ordine as `oi1` from `i1`) `ti1`,
    (select valore as `vi2`,ordine as `oi2` from `i2`) `ti2`,
    (select valore as `vi3`,ordine as `oi3` from `i3`) `ti3`,
    (select valore as `vi4`,ordine as `oi4` from `i4`) `ti4`
) allkeys

left join

    (select * from `dati`
        where (`i1`= 'TARGET|TOTALE')
        AND (`i2`= 'BANCHE|ADV AWARENESS SPONTANEA OFFLINE')
        AND (`i3`= 'BRAND BANCARI|(NET) BANCOPOSTA')
        AND (`i4`= 'ANNO|2014') AND (TRUE)
    ) core

on ( `allkeys`.`vi1`=`core`.`i1` and `allkeys`.`vi2`=`core`.`i2`and `allkeys`.`vi3`=`core`.`i3` and `allkeys`.`vi4`=`core`.`i4` ) 
)

Answer 1

您可以通过连接实现所有这些，但由于您的数据量很大，您必须索引您的表以获得更快的查询响应。 您可以将索引应用于更频繁检索或在连接中一次又一次使用的列。 还请尝试指定选择性列名而不是 *，因为这样可以更快地处理您的列，而不是从所有列中获取数据。