I'm trying to order a list of locations based on the number of times a user has viewed them. Am using the impressionist gem for the sake of it.
The problem I'm having is that my query completely excludes those locations the user's never viewed. I need to display these at the bottom of the results and order by the created_at timestamp.
I can do this to get a list of location_ids:
#location_ids = #user.impressions.
select('count(id) as counter, impressionable_id').
group(:impressionable_id).
order('counter DESC').
#location_ids.map(&:impressionable_id)
Which gives [3,5,8,44,99] and so on..
However, that doesn't get me far so I tried this:
#user.locations.
joins(:impressions).
select("count(impressions.id) as counter, impressionable_id, locations.location_name, locations.id").
group(:impressionable_id).
order("counter desc")
Which is better but it omits those locations with zero views.
How should I do this to get all the locations?
By default, Rails uses an inner join when you use .joins. That's why you don't see the locations with no associated impressions. You need to tell it to use a left join instead, probably like so:
#user.locations.
joins("left join impressions on impressions.impressionable_id = locations.id and impressions.impressionable_type = 'Location'").
select("count(impressions.id) as counter, impressionable_id, locations.location_name, locations.id").
group('locations.id').
order("counter desc")
Related
Problem:
I am using the ransack gem to sort columns in a table. I have 2 models: Campaign and Course. A campaign has many courses, and a course belongs to one campaign. Each course has a number of total_attendees. My Campaigns table has a column for Total Attendees, and I want it to be sortable. So it would sum up the total_attendees field for each course that belongs to a single campaign, and sort based on that sum.
Ex. A campaign has 3 courses, each with 10 attendees. The Total Attendees column on the campaign table would show 30 and it would be sortable against total attendees for all the other campaigns.
I found ransackers:
https://github.com/activerecord-hackery/ransack/wiki/Using-Ransackers
and this SO question: Ransack sort by sum of relation
and from that put together a lot of what is below.
From Model - campaign.rb:
class Campaign < ApplicationRecord
has_many :courses
ransacker :sum_of_total_attendees do
query = "SELECT SUM(r.total_attendees)
FROM campaigns c
LEFT OUTER JOIN courses r
ON r.campaign_id = c.id
GROUP BY c.id"
Arel.sql(query)
end
end
From Model - course.rb:
class Course < ApplicationRecord
belongs_to :campaign, optional: true
end
View:
<th scope="col"><%= sort_link(#q, :sum_of_total_attendees, 'Total Attendees') %></th>
Controller - campaigns_controller.rb:
all_campaigns = Campaign.all
#q = all_campaigns.ransack(params[:q])
#campaigns = #q.result
Errors:
The ransacker query gives me the data I want, but I don't know what to do to get the right information .
Originally, when I clicked on the th link to sort the data, I got this error:
PG::CardinalityViolation: ERROR: more than one row returned by a
subquery used as an expression
I don't know what changed, but now I'm getting this error:
PG::SyntaxError: ERROR: syntax error at or near "SELECT"
LINE 1: SELECT "campaigns".* FROM "campaigns" ORDER BY SELECT SUM(r....
^
: SELECT "campaigns".* FROM "campaigns" ORDER BY SELECT
SUM(r.total_attendees)
FROM campaigns c
LEFT OUTER JOIN courses r
ON r.campaign_id = c.id
GROUP BY c.id ASC
This error seems to say that the ransack search parameter, #q and the ransacker query don't work together. There are two selects in this request, when there should definitely be only one, but the first one is coming from ransack, so I'm not sure how to address it.
How do I get my query to sort correctly with ransack?
Articles I've looked at but did not seem to apply to what I was looking to accomplish with this story:
Ransack Sort By Sum of Relation: This is the one I worked from a lot, but I'm not sure why it works for this user and not for me. They don't show what is changed, if anything, in the controller
Ransack Github Issue For Multiple Params: This doesn't cover the issue of summing table columns.
Rails Ransack Sorting Searching Based On A Definition In The Model: This didn't apply to my need to sort based on summed data.
Three Ways to Bend The Ransack Gem: This looks like what I was doing, but I'm not sure why theirs is working but mine isn't.
In my rails app, I have Users and Listings. The Listings belong to a User. Listing has user_id and its filled with users id who is creating the listing.
A user can be a premium user, gold user or silver user.
What I want is for each premium user, select one random listing to show in premium listings.
I can do it in O(n**2) time or n+1 query as follow:
users_id = User.where(:role => "premium").pluck[:id]
final_array = Array.new
users_id.each do |id|
final_array << Listing.where(:user_id => id).sample(1)
end
final_array
Is there a better way of doing this?
You could try this:
listings = Listing.select(
<<~SQL
DISTINCT ON (users.id) users.id,
listings.*,
row_number() OVER (PARTITION BY users.id ORDER BY random())
SQL
)
.joins(:user)
.includes(:user)
.where(users: { role: :premium })
It gives a random Listing for every premium user.
It produces the only request to db and also it won't make an extra request for getting listing's user, so you are free to do something like this:
listings.each do |listing|
p listing.user
end
random_user_listings = []
User.includes(:listings).where(role: "premium").find_each do |user|
random_user_listings << user.listings.sample(1)
end
random_user_listings
To avoid N+1 query you need to combine them, perform query one time like this:
list = Listing.includes(:user).where(:role => "premium").sample(1)
Feel free to deal with list instead of Listing. Because now you're dealing with variable, not Query.
ids = list.pluck(:user_id).uniq
Getting array of ids like above and doing further steps as you did (but with list, not Listing)
Need to be noticed that, when you deal with Model you're dealing with QUERY. Avoiding doing that in loop statement.
I have two models
class User
has_many :subscriptions
end
and
class Subscription
belongs_to :user
end
one one of my pages I would like to display a list of all users ordered by the number of subscriptions each user has. I am not to good with sql queries but I think that
list = Users.all.joins(:subscriptions).group("user.id").order("count(subscriptions.id) DESC")
dose the job. Now to my problem, when I try to count the total number of objects in list, using list.count, I get a hash with user.id and subscription count, like this
{11 => 5,
8 => 7,
1 => 11,
...}
not the total number of users in list.. .count works fine if I have a list sorted by for example user name (which is in the user table). I would really like to use .count since it in a module for pagination thats in a gem but any ideas is great!
Thanks!
We can just use a single query to finish this:
User.joins("LEFT OUTER JOIN ( SELECT user_id, COUNT(*) as num_subscriptions
FROM subscriptions
GROUP BY user_id
) AS temp
ON temp.user_id = users.id")
.order("temp.num_subscriptions DESC")
Basically, my idea is to try to query the number of subscription for each user_id in the subquery, then join with User. I used LEFT OUTER JOIN, because there will be several users which don't have any subscriptions
Improve option: You can define a scope inside User, it would be more beautiful for later usage:
user.rb
class User < ActiveRecord::Base
has_many :subscriptions
scope :sorted_by_num_subscriptions, -> {
joins("LEFT OUTER JOIN ( SELECT user_id, COUNT(*) as num_subscriptions
FROM subscriptions
GROUP BY user_id
) AS temp
ON temp.user_id = users.id")
.order("temp.num_subscriptions DESC")
}
end
Then just use it:
User.sorted_by_num_subscriptions
When grouping, the count method changes it's behavior and indeed, instead of returning the total count of records, it returns a hash of the counts for each group (see the docs for more info). So what you get with list.count is simply a hash of the subscription counts for each user.
So, your query is correct and all you need is to sum up the individual counts in the groups. This can be done simply by:
total_count = list.count.values.sum
If it is the pagination code that calls just a bare count that makes the issue, usually the pagination code is able to accept a parameter with total count. For example, will_paginate accepts the total_entries parameter, so you should be able to pass it the total count like this:
list.paginate(page: 2, total_entries: list.count.values.sum)
I have a mildly-complex ActiveRecord query in Rails 3.2 / Postgres that returns documents that are related and most relevant to all documents a user has favorited in the past.
The problem is that despite specifying uniq my query does not return distinct document records:
Document.joins("INNER JOIN related_documents ON
documents.docid = related_documents.docid_id")
.select("documents.*, related_documents.relevance_score")
.where("related_documents.document_id IN (?)",
some_user.favorited_documents)
.order("related_documents.relevance_score DESC")
.uniq
.limit(10)
I use a RelatedDocument join table, ranking each relation by a related_document.relevance_score which I use to order the query result before sampling the top 10. (See this question for schema description.)
The problem is that because I select("documents.*, related_documents.relevance_score"), the same document record returned multiple times with different relevance_scores are considered unique results. (i.e. if the document is a related_document for multiple favorited-documents.)
How do I return unique Documents regardless of the related_document.relevance_score?
I have tried splitting the select into two seperate selects, and changing the position of uniq in the query with no success.
Unfortunately I must select("related_documents.relevance_score") so as to order the results by this field.
Thanks!
UPDATE - SOLUTION
Thanks to Jethroo below, GROUP BY is the needed addition, giving me the follow working query:
Document.joins("INNER JOIN related_documents ON
documents.docid = related_documents.docid_id")
.select("documents.*, max(related_documents.relevance_score)")
.where("related_documents.document_id IN (?)",
some_user.favorited_documents)
.order("related_documents.relevance_score DESC")
.group("documents.id")
.uniq
.limit(10)
Have you tried to group it by documents.docid see http://guides.rubyonrails.org/active_record_querying.html#group?
My customer model has many videos and videos has many video activities.
I want to join on video activities to limit based off videos that belong to a customer who has a specific email domain.
This code will give me all the video activities belonging to customer with id 52, but since videos don't have customer email, I need to join customer onto video and then do a .where.
VideoActivity.joins(:video).where(videos: {customer_id: 52})
How is this done? Doing VideoActivity.joins(:video).joins(:customer) gives me an error saying VideoActivity doesn't have a customer associated with it.
VideoActiviy has no relation with customer, you need to say that the customer is related to the video, then it's easier to use active record's #merge that doing a hash where
VideoActivity.joins(video: :customer).merge(Customer.where('some condition'))
If you have a scope in videos you could use that too, here's an example
VideoActivity.joins(video: :customer).merge(Customer.some_scope)
PS:
a scope could be
# customer model
scope :email_ends_with, ->(string) { where('email ilike ?', "%#{string}") }
Then use it
VideoActivity.joins(video: :customer).merge(Customer.email_ends_with('gmail.com'))
There are a bunch of ways to do this and all end up at about the same place, but using an explicit where statement will easily accomplish this goal.
VideoActivity.joins(:video).where("videos.customer_id = ?", 52)
You can write plain sql to fetch records
activities = VideoActivity.joins('INNER JOIN videos ON video_activities.video_id = videos.id INNER JOIN customers ON videos.customer_id = customers.id')
activities.where('customers.email = ?', customer_email)