As maintainer of both Acts As List and the Ranked Model gems, I've become intimately aquainted with the strengths and weaknesses of each. I ended up writing a small scale Rails Concern for positioning database rows for a recent project and it worked really well so I've decided to release it as a gem: Positioning
Positioning works similarly to Acts As List in that it maintains a sequential list of integer values as positions. It differs in that it encourages a unique constraints on the position column and supports multiple lists per database table. It borrows Ranked Model's concept of relative positioning. I encourage you to check it out and give it a whirl on your project!
This acts_as
extension provides the capabilities for sorting and reordering a number of objects in a list. The class that has this specified needs to have a position
column defined as an integer on the mapped database table.
There are a couple of changes of behaviour from 0.8.0
onwards:
- If you specify
add_new_at: :top
, new items will be added to the top of the list like always. But now, if you specify a position at insert time:.create(position: 3)
, the position will be respected. In this example, the item will end up at position3
and will move other items further down the list. Before0.8.0
the position would be ignored and the item would still be added to the top of the list. #220 acts_as_list
now copes with disparate position integers (i.e. gaps between the numbers). There has been a change in behaviour for thehigher_items
method. It now returns items with the first item in the collection being the closest item to the reference item, and the last item in the collection being the furthest from the reference item (a.k.a. the first item in the list). #223
In your Gemfile:
gem 'acts_as_list'
Or, from the command line:
gem install acts_as_list
At first, you need to add a position
column to desired table:
rails g migration AddPositionToTodoItem position:integer
rake db:migrate
After that you can use acts_as_list
method in the model:
class TodoList < ActiveRecord::Base
has_many :todo_items, -> { order(position: :asc) }
end
class TodoItem < ActiveRecord::Base
belongs_to :todo_list
acts_as_list scope: :todo_list
end
todo_list = TodoList.find(...)
todo_list.todo_items.first.move_to_bottom
todo_list.todo_items.last.move_higher
You'll have a number of methods added to each instance of the ActiveRecord model that to which acts_as_list
is added.
In acts_as_list
, "higher" means further up the list (a lower position
), and "lower" means further down the list (a higher position
). That can be confusing, so it might make sense to add tests that validate that you're using the right method given your context.
list_item.insert_at(2)
list_item.move_lower
will do nothing if the item is the lowest itemlist_item.move_higher
will do nothing if the item is the highest itemlist_item.move_to_bottom
list_item.move_to_top
list_item.remove_from_list
list_item.increment_position
list_item.decrement_position
list_item.set_list_position(3)
list_item.first?
list_item.last?
list_item.in_list?
list_item.not_in_list?
list_item.default_position?
list_item.higher_item
list_item.higher_items
will return all the items abovelist_item
in the list (ordered by the position, ascending)list_item.lower_item
list_item.lower_items
will return all the items belowlist_item
in the list (ordered by the position, ascending)
As it stands acts_as_list
requires position values to be set on the model before the instance methods above will work. Adding something like the below to your migration will set the default position. Change the parameters to order if you want a different initial ordering.
class AddPositionToTodoItem < ActiveRecord::Migration
def change
add_column :todo_items, :position, :integer
TodoItem.order(:updated_at).each.with_index(1) do |todo_item, index|
todo_item.update_column :position, index
end
end
end
If you are using the scope option things can get a bit more complicated. Let's say you have acts_as_list scope: :todo_list
, you might instead need something like this:
TodoList.all.each do |todo_list|
todo_list.todo_items.order(:updated_at).each.with_index(1) do |todo_item, index|
todo_item.update_column :position, index
end
end
When using PostgreSQL, it is also possible to leave this migration up to the database layer. Inside of the change
block you could write:
execute <<~SQL.squish
UPDATE todo_items
SET position = mapping.new_position
FROM (
SELECT
id,
ROW_NUMBER() OVER (
PARTITION BY todo_list_id
ORDER BY updated_at
) AS new_position
FROM todo_items
) AS mapping
WHERE todo_items.id = mapping.id;
SQL
All position
queries (select, update, etc.) inside gem methods are executed without the default scope (i.e. Model.unscoped
), this will prevent nasty issues when the default scope is different from acts_as_list
scope.
The position
column is set after validations are called, so you should not put a presence
validation on the position
column.
If you need a scope by a non-association field you should pass an array, containing field name, to a scope:
class TodoItem < ActiveRecord::Base
# `task_category` is a plain text field (e.g. 'work', 'shopping', 'meeting'), not an association
acts_as_list scope: [:task_category]
end
You can also add multiple scopes in this fashion:
class TodoItem < ActiveRecord::Base
belongs_to :todo_list
acts_as_list scope: [:task_category, :todo_list_id]
end
Furthermore, you can optionally include a hash of fixed parameters that will be included in all queries:
class TodoItem < ActiveRecord::Base
belongs_to :todo_list
# or `discarded_at` if using discard
acts_as_list scope: [:task_category, :todo_list_id, deleted_at: nil]
end
This is useful when using this gem in conjunction with the popular acts_as_paranoid or discard gems.
column
default:position
. Use this option if the column name in your database is different from position.top_of_list
default:1
. Use this option to define the top of the list. Use 0 to make the collection act more like an array in its indexing.add_new_at
default::bottom
. Use this option to specify whether objects get added to the:top
or:bottom
of the list.nil
will result in new items not being added to the list on create, i.e, position will be kept nil after create.touch_on_update
default:true
. Usetouch_on_update: false
if you don't want to update the timestamps of the associated records.sequential_updates
Specifies whether insert_at should update objects positions during shuffling one by one to respect position column unique not null constraint. Defaults to true if position column has unique index, otherwise false. If constraint isdeferrable initially deferred
(PostgreSQL), overriding it with false will speed up insert_at.
If you need to temporarily disable acts_as_list
during specific operations such as mass-update or imports:
TodoItem.acts_as_list_no_update do
perform_mass_update
end
In an acts_as_list_no_update
block, all callbacks are disabled, and positions are not updated. New records will be created with
the default value from the database. It is your responsibility to correctly manage positions
values.
You can also pass an array of classes as an argument to disable database updates on just those classes. It can be any ActiveRecord class that has acts_as_list enabled.
class TodoList < ActiveRecord::Base
has_many :todo_items, -> { order(position: :asc) }
acts_as_list
end
class TodoItem < ActiveRecord::Base
belongs_to :todo_list
has_many :todo_attachments, -> { order(position: :asc) }
acts_as_list scope: :todo_list
end
class TodoAttachment < ActiveRecord::Base
belongs_to :todo_item
acts_as_list scope: :todo_item
end
TodoItem.acts_as_list_no_update([TodoAttachment]) do
TodoItem.find(10).update(position: 2)
TodoAttachment.find(10).update(position: 1)
TodoAttachment.find(11).update(position: 2)
TodoList.find(2).update(position: 3) # For this instance the callbacks will be called because we haven't passed the class as an argument
end
When using this gem in an app with a high amount of concurrency, you may see "deadlock" errors raised by your database server. It's difficult for the gem to provide a solution that fits every app. Here are some steps you can take to mitigate and handle these kinds of errors.
One easy way to reduce deadlocks is to use the most concise gem API available for what you want to accomplish. In this specific example, the more concise API for creating a list item at a position results in one transaction instead of two, and it issues fewer SQL statements. Issuing fewer statements tends to lead to faster transactions. Faster transactions are less likely to deadlock.
Example:
# Good
TodoItem.create(todo_list: todo_list, position: 1)
# Bad
item = TodoItem.create(todo_list: todo_list)
item.insert_at(1)
Deadlocks are always a possibility when updating tables rows concurrently. The general advice from MySQL documentation is to catch these errors and simply retry the transaction; it will probably succeed on another attempt. (see How to Minimize and Handle Deadlocks) Retrying transactions sounds simple, but there are many details that need to be chosen on a per-app basis: How many retry attempts should be made? Should there be a wait time between attempts? What other statements were in the transaction that got rolled back?
Here a simple example of rescuing from deadlock and retrying the operation:
ActiveRecord::Deadlocked
is available in Rails >= 5.1.0.- If you have Rails < 5.1.0, you will need to rescue
ActiveRecord::StatementInvalid
and check#cause
.
attempts_left = 2
while attempts_left > 0
attempts_left -= 1
begin
TodoItem.transaction do
TodoItem.create(todo_list: todo_list, position: 1)
end
attempts_left = 0
rescue ActiveRecord::Deadlocked
raise unless attempts_left > 0
end
end
You can also use the approach suggested in this StackOverflow post: https://stackoverflow.com/questions/4027659/activerecord3-deadlock-retry
In addition to reacting to deadlocks, it is possible to reduce their frequency with more pessimistic locking. This approach uses the parent record as a mutex for the entire list. This kind of locking is very effective at reducing the frequency of deadlocks while updating list items. However, there are some things to keep in mind:
- This locking pattern needs to be used around every call that modifies the list; even if it does not reorder list items.
- This locking pattern effectively serializes operations on the list. The throughput of operations on the list will decrease.
- Locking the parent record may lead to deadlock elsewhere if some other code also locks the parent table.
Example:
todo_list = TodoList.create(name: "The List")
todo_list.with_lock do
item = TodoItem.create(description: "Buy Groceries", todo_list: todo_list, position: 1)
end
Version 0.9.0
adds acts_as_list_no_update
(brendon#244) and compatibility with not-null and uniqueness constraints on the database (brendon#246). These additions shouldn't break compatibility with existing implementations.
As of version 0.7.5
Rails 5 is supported.
All versions 0.1.5
onwards require Rails 3.0.x and higher.
We often hear complaints that position
values are repeated, incorrect etc. For example, #254. To ensure data integrity, you should rely on your database. There are two things you can do:
- Use constraints. If you model
Item
thatbelongs_to
anOrder
, and it has aposition
column, then add a unique constraint onitems
with[:order_id, :position]
. Think of it as a list invariant. What are the properties of your list that don't change no matter how many items you have in it? One such propery is that each item has a distinct position. Another could be that position is always greater than 0. It is strongly recommended that you rely on your database to enforce these invariants or constraints. Here are the docs for PostgreSQL and MySQL. - Use mutexes or row level locks. At its heart the duplicate problem is that of handling concurrency. Adding a contention resolution mechanism like locks will solve it to some extent. But it is not a solution or replacement for constraints. Locks are also prone to deadlocks.
As a library, acts_as_list
may not always have all the context needed to apply these tools. They are much better suited at the application level.
- Sort based feature
- Check out the latest master to make sure the feature hasn't been implemented or the bug hasn't been fixed yet
- Check out the issue tracker to make sure someone already hasn't requested it and/or contributed it
- Fork the project
- Start a feature/bugfix branch
- Commit and push until you are happy with your contribution
- Make sure to add tests for it. This is important so I don't break it in a future version unintentionally.
- Please try not to mess with the Rakefile, version, or history. If you want to have your own version, or is otherwise necessary, that is fine, but please isolate to its own commit so I can cherry-pick around it.
- I would recommend using Rails 3.1.x and higher for testing the build before a pull request. The current test harness does not quite work with 3.0.x. The plugin itself works, but the issue lies with testing infrastructure.
Copyright (c) 2007 David Heinemeier Hansson, released under the MIT license