Question

我有一个大的评估表，在该表中有几个带有数字分数的整数列。我试图删除一些N + 1迭代来加速我的应用程序。

我可以在原始SQL中生成此查询以收集所有分数

query = 
  'SELECT ' + 
  Assessment.rating_attributes.collect{|attribute|
    1.upto(5).collect do |i|
      %Q{SUM(CASE WHEN #{attribute.to_s} = #{i} then 1 else 0 end) as #{attribute}_score_#{i}}
    end
    }.join(', ') + ' FROM assessments'

当我用：

执行此查询时

  $> ActiveRecord::Base.connection.execute(query)
 => #<PG::Result:0x007fc9e7e541e8 status=PGRES_TUPLES_OK ntuples=1 nfields=55 cmd_tuples=1>

我得到了正确的结果：

 ActiveRecord::Base.connection.execute(query).first.values
 => ["2400", "458", "3343", "1021", "93", "2352", "455", "3119", "1174", "150", "2378", "357", "2976", "1357", "181", "2042", "258", "3024", "1646", "280", "1274", "193", "3907", "1704", "172", "799", "93", "4593", "1670", "95", "1759", "361", "2542", "947", "69", "21", "100", "81", "42", "38", "8", "32", "51", "78", "112", "19", "119", "72", "43", "29", "0", "0", "0", "0", "0"]

我想执行相同的查询，但不是整个表。我希望能够将其链接到activerecord查询中，例如：

Assessment.where(selection_attribute: 1038).select(query_without_select_keyword_or_from_clause)

ActiveRecord::StatementInvalid: PG::GroupingError: ERROR:  column "assessments.id" must appear in the GROUP BY clause or be used in an aggregate function

让它给我相同的结果，但在子集上。我天真的努力（例如我的例子中的.select（查询）调用）失败了，我甚至不确定如何处理这个问题。

我已经碰壁了，我很欣赏任何指导，甚至关于如何寻找解决方案。谢谢。

Answer 1

嗯，这很简单。使用.first是个问题。我想这是因为查询没有返回模型记录。如果我直接解决结果的0索引，我会得到所有分数。

所以这段代码有效：

 query = 
      Assessment.rating_attributes.collect{|attribute|
        1.upto(5).collect do |i|
        %Q{sum(CASE WHEN #{attribute.to_s} = #{i} then 1 else 0 end) as #{attribute}_score_#{i}}
      end
    }.join(', ')

Assessment.where(selection_attribute: 1038).select(query)[0]
=> #<Assessment id: nil>

Assessment.where(selection_attribute: 1038).select(query)[0].first_attribute_score_1
=> 3

在ActiveRecord查询中包含自定义SELECT子句

1 个答案: