我创建了下表，其中包含有关产品的信息：
CREATE TABLE `AllProducts` (
`insertionTime` datetime NOT NULL,
`shop` varchar(40) CHARACTER SET utf8 NOT NULL,
`name` varchar(100) CHARACTER SET utf8 NOT NULL,
`category` varchar(40) CHARACTER SET utf8 NOT NULL,
`description` text CHARACTER SET utf8 NOT NULL,
`price` decimal(10,2) NOT NULL,
`url` varchar(350) COLLATE latin1_german1_ci NOT NULL,
`image` varchar(450) CHARACTER SET utf8 NOT NULL,
`fromPrice` text CHARACTER SET utf8 NOT NULL,
`deliveryCosts` decimal(10,2) NOT NULL,
`stock` text CHARACTER SET utf8 NOT NULL,
`deliveryTime` varchar(80) CHARACTER SET utf8 NOT NULL,
`ean` varchar(32) CHARACTER SET utf8 NOT NULL,
`color` varchar(40) CHARACTER SET utf8 NOT NULL,
`size` varchar(40) CHARACTER SET utf8 NOT NULL,
`brand` varchar(40) CHARACTER SET utf8 NOT NULL,
KEY `category` (`category`),
KEY `nameShopInsertiontime` (`name`,`shop`,`insertionTime`),
KEY `name` (`name`),
KEY `catNameShopInsert` (`category`,`name`,`shop`,`insertionTime`),
KEY `eanName` (`ean`,`name`),
KEY `eanInsertionTime` (`ean`,`insertionTime`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 COLLATE=latin1_german1_c
我想创建一个从该表中选择所有独特产品的视图（“原型产品”视图）。我表中的一些产品有 EAN 编号，但不是全部。如果有可用的 EAN 编号，我想将其用作 productId。如果它不可用，我想散列名称（小写）并将其用作产品 ID。
我创建了以下视图，它使用联合所有来区分上述情况。它可以工作，但需要 4.3 秒才能完成（在 200.000 行上）。这太慢了。
select
max(`ap`.`name`) AS `name`,
max(`ap`.`description`) AS `description`,
min(`ap`.`price`) AS `price`,
max(`ap`.`image`) AS `image`,
`ap`.`ean` AS `ean`,
max(`ap`.`color`) AS `color`,
max(`ap`.`size`) AS `size`,
max(`ap`.`brand`) AS `brand`,
`ap`.`ean` AS `productId`
from `AllProducts` `ap`
where (not((`ap`.`ean` like '')))
group by `ap`.`ean`
union all
select
`ap`.`name` AS `name`,
max(`ap`.`description`) AS `description`,
min(`ap`.`price`) AS `price`,
max(`ap`.`image`) AS `image`,
'' AS `ean`,
max(`ap`.`color`) AS `color`,
max(`ap`.`size`) AS `size`,
max(`ap`.`brand`) AS `brand`,
md5(lcase(`ap`.`name`)) AS `productId`
from `AllProducts` `ap`
where (`ap`.`ean` = '')
group by `ap`.`name` order by NULL
如果有多种产品可用，则需要最长/最短的可用描述（或颜色/尺寸/其他）。我不一定需要最长或最短的值，而只需要任何值。
我的问题：我能做些什么来提高这个视图的性能？
使用 Mani 的评论，我切换到以下情况，这确实很快。
select
`ap`.`name` AS `name`,
`ap`.`description` AS `description`,
a.category,
`ap`.`price` AS `price`,
`ap`.`image` AS `image`,
`ap`.`ean` AS `ean`,
`ap`.`color` AS `color`,
`ap`.`size` AS `size`,
`ap`.`brand` AS `brand`,
case when ap.ean not like '' then ap.ean else md5(lcase(ap.name)) end as productId
from `AllProducts` `ap`
这个视图是相当快的。但是，我仍然需要从这个新视图中选择唯一的 productId——包括名称、描述、类别等（视图称为 AllProductsView）。我尝试了以下方法，这导致查询比原始查询慢（6.5 秒）：
select max(`apv`.`name`) AS `name`,
max(`apv`.`description`) AS `description`,
max(`apv`.`category`) AS `category`,
min(`apv`.`price`) AS `price`,
max(`apv`.`image`) AS `image`,
max(`apv`.`color`) AS `color`,
max(`apv`.`size`) AS `size`,
max(`apv`.`brand`) AS `brand`,
`apv`.`productId` AS `productId`
from `AllProductsView` `apv`
group by `apv`.`productId`
任何人有建议如何加快这个查询？