I need to do something similar to this (see search filters on the left) http://www.indeed.com/job开发者_Go百科s?q=php&l=NY
I've already done everything I know and even though it became much faster than it used to, it's still not fast enough. I have many further optimization ideas but they are basically blind shots, I don't know whether that will work. So please check my current code and let me know if there are some serious mistake or some place that I overlooked. Below is the structure of the table, it has approximately 100k entries and is pretty huge - 640 MiB according to phpmyadmin. Also below is one of the queries I use to gather data for the filters, I won't copy other queries here because they are practically identical. The query time varies from 0.5 to 1 second which sum ups in approximately 5 seconds after I execute a query for each filter I need.
CREATE TABLE `jobs` (
`id` int(11) NOT NULL auto_increment,
`employer_id` int(11) NOT NULL default '0',
`job_title` varchar(255) NOT NULL default '',
`city` varchar(96) NOT NULL default '',
`state` varchar(128) NOT NULL default '0',
`country` varchar(48) NOT NULL default '0',
`zip` varchar(10) NOT NULL default '0',
`zlat` double NOT NULL default '0',
`zlong` double NOT NULL default '0',
`job_type` varchar(255) NOT NULL default '0',
`position_type` varchar(255) NOT NULL default '0',
`req_education` varchar(255) NOT NULL default '0',
`req_experience` varchar(255) NOT NULL default '0',
`salary_type` varchar(255) default '0',
`approx_salary` decimal(10,2) NOT NULL default '0.00',
`approx_salary_range` decimal(10,2) NOT NULL default '0.00',
`short_description` varchar(255) NOT NULL default '',
`long_description` text NOT NULL,
`pay_low` decimal(10,2) NOT NULL default '0.00',
`pay_high` decimal(10,2) NOT NULL default '0.00',
`step_completed` tinyint(1) NOT NULL default '0',
`num_viewed` int(11) NOT NULL default '0',
`num_applicants` int(11) NOT NULL default '0',
`time` int(11) NOT NULL default '0',
`end_time` int(11) NOT NULL default '0',
`status` varchar(15) NOT NULL default 'active',
`service_type` int(11) NOT NULL default '0',
`req_travel` varchar(255) NOT NULL default '',
`tele_position` varchar(255) NOT NULL default '',
`manage_app` varchar(255) NOT NULL default '',
`apply_email_url` varchar(255) NOT NULL default '',
`adcourier_email` text NOT NULL,
`adcourier_url` text NOT NULL,
`reference_number` varchar(255) NOT NULL,
`feed_id` int(11) NOT NULL,
`mods_feed_company_name` varchar(255) NOT NULL COMMENT 'A value in this column automaticaly means that users.is_bot=1 for this employer',
`current_company_name` varchar(255) NOT NULL COMMENT 'this field lets us avoid JOINs in search queries',
`salary_total_grouped` int(11) NOT NULL COMMENT 'a calculated field that is used to speed up search',
`company_type_grouped` varchar(30) NOT NULL COMMENT 'a calculated field that is used to speed up search',
`location_grouped` varchar(100) NOT NULL COMMENT 'a calculated field that is used to speed up search',
PRIMARY KEY (`id`),
KEY `employer_id` (`employer_id`),
KEY `status` (`status`),
KEY `zip` (`zip`),
KEY `zlat` (`zlat`),
KEY `zlong` (`zlong`),
KEY `feed_id` (`feed_id`),
KEY `service_type` (`service_type`),
KEY `time` (`time`),
KEY `job_title` (`job_title`),
KEY `short_description` (`short_description`),
KEY `current_company_name` (`current_company_name`),
KEY `salary_total_grouped` (`salary_total_grouped`),
KEY `location_grouped` (`location_grouped`),
KEY `company_type_grouped` (`company_type_grouped`),
KEY `position_type` (`position_type`),
KEY `job_type` (`job_type`),
KEY `salary_type` (`salary_type`),
KEY `approx_salary` (`approx_salary`),
KEY `approx_salary_range` (`approx_salary_range`),
KEY `country` (`country`),
KEY `state` (`state`),
KEY `city` (`city`),
KEY `end_time` (`end_time`),
FULLTEXT KEY `long_description` (`long_description`)
) ENGINE=MyISAM AUTO_INCREMENT=485825 DEFAULT CHARSET=latin1
SELECT count( jobs.job_title ) AS count, jobs.job_title
FROM jobs
WHERE jobs.status = 'active'
AND jobs.country = 'United States'
GROUP BY jobs.job_title
ORDER BY count DESC
LIMIT 14
Try add an index on status, country, job_title
, and do execution query plan to exam the performance
Think about the first, second, third and fourth database normal forms: http://en.wikipedia.org/wiki/Database_normalization
精彩评论