原始sql查询和使用活动记录查询的查询写操作给出不同的结果

Question

我正在运行两个查询，在第一个中，我运行原始sql查询：

SELECT  drivers.*, shift_timing.* FROM "drivers" 
        LEFT JOIN(SELECT DISTINCT ON (driver_id) driver_id, 
        start_hour, start_minutes, end_hour, end_minutes FROM 
        shift_timings ORDER BY driver_id) AS shift_timing ON 
        drivers.id = shift_timing.driver_id WHERE 
        "drivers"."deleted_at" IS NULL  ORDER BY "drivers"."id" ASC LIMIT 1

这给了我以下输出：

{"id"=>"1", "driver_name"=>"Narendra Kumar Soni", ...,
 "last_order_id"=>"31550", "last_shift_id"=>"74483", "driver_id"=>"1", 
 "start_hour"=>"7", "start_minutes"=>"0", "end_hour"=>"19", 
 "end_minutes"=>"0"}

但是当我运行以下查询时，它将完全转换为原始sql查询：

Driver.joins('LEFT JOIN(SELECT DISTINCT ON (driver_id) driver_id, 
              start_hour, start_minutes, end_hour, end_minutes FROM 
              shift_timings ORDER BY driver_id) 
              AS shift_timing ON drivers.id = 
              shift_timing.driver_id')
             .select('drivers.*, shift_timing.*').first    

我得到以下结果：

#<Driver id: 1, driver_name: "Narendra Kumar Soni", ...,
, last_order_id: 31550, last_shift_id: 74483>

基本上，活动记录查询结果中缺少start_hour，start_minutes，end_hour，end_minutes列，而缺少的列实际上是排班计时表的列。 有人可以解释为什么会这样吗？如何解决？

Answer 1

通常可以使用['name']语法访问联接表中['name'] ，因此外观类似于driver['start_hour'] 。

但是我认为解决该问题的正确方法只是包括shift_timings表，例如

Driver.includes(:shift_timings) 。

Answer 2

要么删除select子句，以获得完整的ActiveRecord对象和用途includes(:shift_timings)替代，或者用pluck只得到你想要作为一个数组的字段：

Driver.joins('LEFT JOIN ...') # As in the question
      .pluck('drivers.driver_name, shift_timings.start_hour, shift_timings.start_minutes')