ruby-on-rails – 带有group的SQL查询可以在SQLite上运行,而不是在postgresql上运行

以下查询适用于本地但不适用于生产:(生产是heroku正在运行postgreSQL,本地我正在运行sqllite数据库)

ruby

Tutor.joins(:expertises).where(:expertises => {:subject_id => [2,4]}).group("tutors.id").having("COUNT(*) = 2")

SQL

SELECT "tutors".* FROM "tutors" INNER JOIN "expertises" ON "expertises"."tutor_id" = "tutors"."id" WHERE ("expertises"."subject_id" IN (9)) GROUP BY tutors.id HAVING COUNT(*) = 1 ORDER BY rank DESC)

我在生产ActiveRecord :: StatementInvalid时遇到以下错误(PGError：错误：列“tutors.fname”必须出现在GROUP BY子句中或用于聚合函数

我的表中有以下值

id                 :integer         not null, primary key
fname              :string(255)
lname              :string(255)
school             :string(255)
major              :string(255)
year               :string(255)
bio                :text
vid                :string(255)
price              :integer
phone              :string(255)
skype              :string(255)
address            :text

当我尝试将查询调整为按所有属性分组时,我收到另一个错误：

ruby

>> Tutor.joins(:expertises).where(:expertises => {:subject_id => [2,4]}).group("tutors.*").having("COUNT(*) = 2")

SQL

SELECT "tutors".* FROM "tutors" INNER JOIN "expertises" ON "expertises"."tutor_id" =    "tutors"."id" WHERE ("expertises"."subject_id" IN (2, 4)) GROUP BY tutors.* HAVING COUNT(*)    = 2 ORDER BY rank DESC

ActiveRecord :: StatementInvalid：PGError：ERROR：无法识别类型教师的排序运算符HINT：使用显式排序运算符或修改查询.

救命！

嗯,很好的尝试,但星号扩展在GROUP BY子句中不起作用.您需要实际列出所有列.或者在几个月内升级到PostgreSQL 9.1.