MySQL query optimization - increase speed and efficiency - mysql

Question

I want to get some emails in a database and each email have a status. All the possible status are stock in a table where they all have permissions (such as show, edit, delete, etc.). Those emails are not users with permissions trough a site but a list of emails a user have added.
Here is the tables structure:
Email table
CREATE TABLE IF NOT EXISTS `email__email` (
`email_id` int(11) NOT NULL AUTO_INCREMENT,
`created` timestamp NULL DEFAULT NULL,
`updated` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
`user_fk` int(11) NOT NULL,
`status_fk` tinyint(2) NOT NULL,
`language` enum('fr','en') COLLATE utf8_unicode_ci DEFAULT NULL,
`email` varchar(255) COLLATE utf8_unicode_ci DEFAULT NULL,
`firstName` varchar(255) COLLATE utf8_unicode_ci DEFAULT NULL,
`lastName` varchar(100) COLLATE utf8_unicode_ci DEFAULT NULL,
`companyName` varchar(255) COLLATE utf8_unicode_ci DEFAULT NULL,
`gender` enum('f','m') COLLATE utf8_unicode_ci DEFAULT NULL,
PRIMARY KEY (`email_id`),
UNIQUE KEY `user_email` (`user_fk`,`email`),
KEY `user_fk` (`user_fk`),
KEY `created` (`created`),
KEY `status_fk` (`status_fk`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci AUTO_INCREMENT=3031492 ;
Status table
CREATE TABLE IF NOT EXISTS `email__status` (
`status_id` int(11) NOT NULL AUTO_INCREMENT,
`name_fr` varchar(50) COLLATE utf8_unicode_ci DEFAULT NULL,
`name_en` varchar(50) COLLATE utf8_unicode_ci DEFAULT NULL,
`description_fr` varchar(150) COLLATE utf8_unicode_ci DEFAULT NULL,
`description_en` varchar(150) COLLATE utf8_unicode_ci DEFAULT NULL,
`permShow` tinyint(1) NOT NULL DEFAULT '0',
`permSend` tinyint(1) NOT NULL DEFAULT '0',
`permEdit` tinyint(1) NOT NULL DEFAULT '0',
`permDelete` tinyint(1) NOT NULL DEFAULT '0',
`permImport` tinyint(1) NOT NULL DEFAULT '0',
PRIMARY KEY (`status_id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci AUTO_INCREMENT=7 ;
Here is the slow query with the EXPLAIN:
SELECT EE.*, ES.name_fr AS statusName, ES.description_fr AS statusDescription, ES.permShow, ES.permSend, ES.permEdit, ES.permDelete, ES.permImport
, (SELECT GROUP_CONCAT(CONVERT(CONCAT(GC.name, '~', GC.group_id), CHAR(255)) SEPARATOR ',') FROM `group` GC INNER JOIN group_email GEC ON GEC.group_fk = GC.group_id WHERE GEC.email_fk = EE.email_id AND GC.deleted = 0) AS groups
FROM `email__email` EE
INNER JOIN email__status ES ON EE.status_fk = ES.status_id
WHERE 1 = 1
AND EE.user_fk = 54
AND ES.permShow = 1
ORDER BY EE.email_id DESC LIMIT 15
EXTRA ID KEY KEY_LEN POSSIBLE_KEYS REF ROWS SELECT_TYPE TABLE TYPE
Using temporary; Using filesort 1 user_email 4 user_email,user_fk,status_fk const 180681 PRIMARY EE ref
Using where; Using join buffer 1 [empty string] [empty string] PRIMARY [empty string] 6 PRIMARY ES ALL
Using index 2 email_fk 4 group_email,group_fk,email_fk mailing_dev.EE.email_id 1 DEPENDENT SUBQUERY GEC ref
Using where 2 PRIMARY 4 PRIMARY mailing_dev.GEC.group_fk 1 DEPENDENT SUBQUERY GC eq_ref
Here is a fast query with the EXPLAIN:
SELECT EE.*
, (SELECT GROUP_CONCAT(CONVERT(CONCAT(GC.name, '~', GC.group_id), CHAR(255)) SEPARATOR ',') FROM `group` GC INNER JOIN group_email GEC ON GEC.group_fk = GC.group_id WHERE GEC.email_fk = EE.email_id AND GC.deleted = 0) AS groups
FROM `email__email` EE
WHERE 1 = 1
AND EE.user_fk = 54
AND EXISTS(SELECT permShow FROM email__status WHERE status_id = EE.status_fk AND permShow = 1)
ORDER BY EE.email_id DESC LIMIT 15
EXTRA ID KEY KEY_LEN POSSIBLE_KEYS REF ROWS SELECT_TYPE TABLE TYPE
Using where 1 PRIMARY 4 user_email,user_fk [empty string] 270 PRIMARY EE index
Using where 3 PRIMARY 4 PRIMARY mailing_dev.EE.status_fk 1 DEPENDENT SUBQUERY email__status eq_ref
Using index 2 email_fk 4 group_email,group_fk,email_fk mailing_dev.EE.email_id 1 DEPENDENT SUBQUERY GEC ref
Using where 2 PRIMARY 4 PRIMARY mailing_dev.GEC.group_fk 1 DEPENDENT SUBQUERY GC eq_ref
There is a big difference between both queries but the second one doesn't give me two important columns that I need to fetch. I can do subqueries to fetch them like a join would do but still, I don't want to have a lot of subqueries for each... any ideas to improve this ?
Thanks

Answer 1

email__email.status_fk is a tinyint, but email__status.status_id is an int(11).
This probably is fouling up your INNER JOIN. Change one or the other data type and try again.