作者:庄廓然
问题简述
本次开发过程中我们用到了rails的orm框架,使用orm框架可以很方便地进行对象的关联和查询,例如查询一个用户的所有关注的社团可以用一下语句:
list = @user.followed_clubs
#user.rb模型中添加
#user和club是多对多的关系,一个user可以关注多个club,一个club也可以有多个关注者
has_many :user_follow_clubs, dependent: :destroy
has_many :followed_clubs, through: :user_follow_clubs, source: :club
#club.rb模型中添加
has_many :user_follow_clubs, dependent: :destroy
has_many :follow_users, through: :user_follow_clubs, source: :user
但是如果你要返回一个社团列表,并且列表中包含关注该社团的人数,直接调用一下语句会导致查询效率降低
list = @user.followed_clubs
result = []
list.each do |club|
result.append(
{
club_id: club.id,
followers: club.follow_users.count
#该语句会导致在每个club中调用如下sql语句
}
)
end
SELECT COUNT(*) FROM `users` INNER JOIN `user_follow_clubs`
ON `users`.`id` = `user_follow_clubs`.`user_id` WHERE `user_follow_clubs`.`club_id` = xxxx
也就是查询一个社团列表调用了N次额外的查询。
查询主数据,是1次查询,查询出n条记录;根据这n条主记录,查询从记录,共需要n次,所以叫数据库1+n问题
问题解决
group的方法简化查询
list = @user.followed_clubs.ids
count_hash = UserFollowClub.where(club_id: list).group(:club_id).count
SELECT COUNT(*) AS count_all, `user_follow_clubs`.`club_id` AS user_follow_clubs_club_id FROM `user_follow_clubs` WHERE `user_follow_clubs`.`club_id` IN (1033447816, 126833941, 386008940) GROUP BY `user_follow_clubs`.`club_id`
最终得到一个hash,key
对应的是club_id
,value
对应的是关注者的个数
{126833941=>1, 386008940=>2}
没有记录的社团对应的关注者就是0.
所以只用了一条查询记录便找到了所有社团的关注者的数量,提高了查询效率
改正后的代码
list = @user.followed_clubs
id_list = list.ids
count_hash = UserFollowClub.where(club_id: id_list).group(:club_id).count
list.each do |club|
result.append(
{
club_id: club.id,
followers: count_hash[club.id]
}
)
end