Question

我有一张有449383条记录的表。当我执行以下查询时，要花费32643条记录需要很多时间。

以下是查询：

SELECT  alloc_id,  cr_number,  profile_id,  table_id,  provider_id,
        user_id,  team_id,  trans_reason_id,  trans_reason,  data_id,
        trans_fr_provider_id,  trans_to_provider_id,  from_office_team_id,
        to_office_team_id,  trans_fr_resp_officer_id,
        trans_to_resp_officer_id,   fr_team,  to_team,  fr_officer,
        to_officer,  start_date,  end_date,  alloc_date,  request_type,
        tr_id,  desc_id,  desc,  bd_status_id,  active,   display_order,
        created_date,  created_by,  created_by_user_id,  modified_date,
        modified_by,  modified_by_user_id,  deleted,  deleted_date,
        deleted_by,   deleted_by_user_id,  locked,  version
FROM    cm_alloc_spg
WHERE   team_id IS NOT NULL 
    AND end_date IS NULL AND locked = 0 
LIMIT   1000000;

执行计划：

 id  select_type  table         type  possible_keys            key     key_len  ref      rows  Extra   
---  -----------  ------------  ----  -------------            ------  -------  -----  ------  ----------------------------------
  1  SIMPLE       cm_alloc_spg  ref   team_idx,ed_idx,loc_idx  ed_idx       4   const  228838  Using index condition; Using where

创建查询结构：

CREATE TABLE `cm_alloc_spg` (
  `alloc_id` INT(10) NOT NULL AUTO_INCREMENT,
  `cr_number` VARCHAR(50) DEFAULT NULL,
  `profile_id` INT(11) NOT NULL,
  `table_id` INT(11) NOT NULL,
  `provider_id` INT(10) DEFAULT NULL,
  `user_id` INT(11) DEFAULT NULL,
  `team_id` INT(11) DEFAULT NULL,
  `trans_reason_id` INT(11) DEFAULT NULL,
  `trans_reason` VARCHAR(400) DEFAULT NULL,
  `data_id` INT(11) NOT NULL,
  `trans_fr_provider_id` INT(11) DEFAULT NULL,
  `trans_to_provider_id` INT(11) DEFAULT NULL,
  `from_office_team_id` INT(10) DEFAULT NULL,
  `to_office_team_id` INT(10) DEFAULT NULL,
  `trans_fr_resp_officer_id` INT(10) DEFAULT NULL,
  `trans_to_resp_officer_id` INT(10) DEFAULT NULL,
  `fr_team` VARCHAR(100) DEFAULT NULL,
  `to_team` VARCHAR(100) DEFAULT NULL,
  `fr_officer` VARCHAR(100) DEFAULT NULL,
  `to_officer` VARCHAR(100) DEFAULT NULL,
  `start_date` DATE NOT NULL,
  `end_date` DATE DEFAULT NULL,
  `alloc_date` DATE DEFAULT NULL,
  `request_type` VARCHAR(40) DEFAULT NULL,
  `tr_id` INT(11) DEFAULT NULL,
  `desc_id` INT(10) DEFAULT NULL,
  `desc` VARCHAR(700) DEFAULT NULL,
  `bd_status_id` INT(11) DEFAULT NULL,
  `active` TINYINT(4) DEFAULT 0,
  `display_order` INT(4) NOT NULL,
  `created_date` TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP(),
  `created_by` VARCHAR(100) NOT NULL,
  `created_by_user_id` INT(11) NOT NULL,
  `modified_date` TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP(),
  `modified_by` VARCHAR(100) NOT NULL,
  `modified_by_user_id` INT(11) NOT NULL,
  `deleted` TINYINT(4) NOT NULL DEFAULT 0,
  `deleted_date` TIMESTAMP NULL DEFAULT NULL,
  `deleted_by` VARCHAR(100) DEFAULT NULL,
  `deleted_by_user_id` INT(11) DEFAULT NULL,
  `locked` TINYINT(4) NOT NULL DEFAULT 0,
  `version` INT(11) NOT NULL DEFAULT 0,
  PRIMARY KEY (`alloc_id`),
  KEY `fk_profile_id_idx` (`profile_id`),
  KEY `fk_user_id_idx` (`user_id`),
  KEY `team_idx` (`team_id`),
  KEY `cr_number_idx` (`cr_number`),
  KEY `tab_idx` (`table_id`),
  KEY `ed_idx` (`end_date`),
  KEY `loc_idx` (`locked`),
  KEY `alloc_date_idx` (`alloc_date`),
  KEY `tr_id_idx` (`tr_id`),
  KEY `del_idx` (`deleted`),
  KEY `to_off_team_id_idx` (`to_office_team_id`),
  KEY `trns_res_offcr_id_idx` (`trans_to_resp_officer_id`),
  KEY `active_idx` (`active`)
) ENGINE=INNODB AUTO_INCREMENT=1 DEFAULT CHARSET=utf8

如何提高查询效率？

Answer 1

我认为您应该使用分页，如果您有大量记录，建议不要获取所有数据。

SELECT alloc_id，cr_number，profile_id，table_id，provider_id，user_id，team_id，trans_reason_id，trans_reason，data_id， trans_fr_provider_id，trans_to_provider_id，from_office_team_id，to_office_team_id，trans_fr_resp_nzl_id，trans_to_resp_nzl_id，fr_team，to_team，fr_leg，to_leg，start_date，end_date，alloc_date，request_type，by_id，by_id，dested_ed，dested_stat，stat_id，dested_stat，stat_id，dested_stat，stat_id，dested_stat，stat_id， Modify_by_user_id，deleted，deleted_date，delete_by，deleted_by_user_id，锁定，版本FROM cm_alloc_spg WHERE team_id IS NOT NULL AND end_date IS NULL AND lock = 0 LIMIT 0,100;

limit子句语法索引，没有要检索的行

Answer 2

古普塔，您只需要在表中创建team_id，end_data并锁定 index 即可。您可以通过以下方式轻松创建任何属性索引例子

创建索引team_id_1 ON cm_alloc_spg（team_id）;

它将使您的查询速度提高30倍以上。注释掉任何帮助。

慢查询-如何提高mysql中的查询性能

2 个答案: