I'm trying to order by two columns, one is an association column which is being ignored, not implemented. Is it possible to order by association in a loop?
Loop
<% @info.each do |data| %> <% data.selections.size.times do |i| %> <% unless data.progressions.blank? %> <% data.progressions.order_scope.in_groups_of(7)[i].each_with_index do |e, index| %> <% if index == 0 %> <%= e.selection.name rescue 0 %> <% end %> <%= Time.at(e.assessment.date/1000).strftime("%d/%m/%Y") rescue 0 %> <% end %> <% end %> <% end %> <% end %> Model
belongs_to :selection, primary_key: 'selection_id', foreign_key: 'selection_id', optional: true belongs_to :assessment, primary_key: 'assessment_id', optional: true scope :order_scope, -> { order('selection_id').joins(:assessment).order('assessments.date') } This is the output the previous code outputs, which seems to be randomly placing the dates instead of in order.
|name 1| 0 |0|0|0|02/01/2018|15/01/2018|23/01/2018| |name 2|28/02/2018|0|0|0|02/01/2018|15/01/2018|21/01/2018| 2 Answers
Answers 1
I think it is because progressions could have the same selection_id,when you order them, the result can be different. Try adding another unique attribute/column into the order, for example created_at or id, so that the order is always unique.
scope :order_scope, -> { order('selection_id, created_at').joins(:assessment).order('assessments.date') } Answers 2
So, what you're trying to do is basically order by 2 columns and use the result to output something? Yes, it can be done. Thing is, I would recommend using SQL instead, because in Ruby methods make it much less readable.
Execute this query in ruby:
SELECT * FROM selections s JOIN assessments a ON s.selection_id = a.selection_id ORDER BY a.date [DESC], s.selection_id [DESC]
[DESC] is optional keyword setting order in descending manner (ascending is default).
Result should be what you're looking for.
0 comments:
Post a Comment