shahryarjb
Scrivener lib doesn't load some records
Hello, I created an issue in scrivener github, but scrivener owner closed my topic without any help or comment. but I still have a problem with this lib
my Topic:
Hello, when I use Repo.all(query) to load 49 records I have the record concerned and I have no problem all records are displayed , but when I use scrivener to load 20 records on each page like this:
def show_error_brands(pagenumber, category_id) do
query = from u in ErrorBrandSchema,
where: u.category_id == ^category_id,
order_by: [desc: u.inserted_at],
join: c in assoc(u, :error_categories),
select: %{
id: u.id,
title: u.title,
description: u.description,
image: u.image,
status: u.status,
category_id: u.category_id,
inserted_at: u.inserted_at,
updated_at: u.updated_at,
category_name: c.title
}
Repo.paginate(query, %{page: pagenumber, page_size: 20})
end
some data is not displayed, but they exist in my database, I use UUID
I use psql 11 and {:scrivener_ecto, "~> 2.2"} , elixir 1.9 , {:phoenix_ecto, "~> 4.0"}
Help me please
my migration:
defmodule BankError.Repo.Migrations.ErrorBrands do
use Ecto.Migration
def change do
create table(:error_brands, primary_key: false) do
add :id, :uuid, primary_key: true
add :title, :string, size: 150, null: false
add :short_description, :string, size: 164, null: false
add :description, :text, null: false
add :image, :string, size: 200, null: false
add :seo_alias_link, :string, size: 200, null: false
add :seo_words, :string, size: 150, null: false
add :seo_description, :string, size: 164, null: false
add :status, :boolean, null: false
add :category_id, references(:error_categories, on_delete: :nothing, type: :uuid)
timestamps()
end
create(
index(:error_brands, [:seo_alias_link],
# concurrently: true,
name: :unique_index_on_error_brands_seo_alias_link,
unique: true
)
)
end
end
Please help me to fix this, Thanks
my issue: scrivener dosent load some records · Issue #83 · mojotech/scrivener_ecto · GitHub
the other issue without any comment: Paginate join incorrect · Issue #61 · mojotech/scrivener_ecto · GitHub
Marked As Solved
NobbZ
Only if they have the same updated_at and id. And I hope id will not clash ever ![]()
Adding the id to the sorted_by is a trick to stabilize the usually unstable sort.
Any column that is guaranteed to have unique values among the result set is fine. For id its easy to guarantee and therefore its used quite often.
Also Liked
NobbZ
You should sort the other way round.
This will only consider updated_at when the ids are equal, which is unlikely to happen…
LostKobrakai
Try adding ordering by your uuid. If rows have the same inserted_at value ordering between them is not deterministic.
NobbZ
What I meant was, you should do order_by: [desc: u.updated_at, desc: u.id].
The problem with only sorting over u.updated_at is, that items with the same updated_at will be ordered randomly each time the query is sent to the DB. This is by the design of most databases. That might cause, that an item close to pagination border to hop between pages, and beeing either displayed on neither page or on both.
This might then be observed as “items appear twice” or “items are missing”.
Popular in Questions
Other popular topics
Categories:
Sub Categories:
Forums
Popular Tags
- #ecto
- #liveview
- #troubleshooting
- #learning-elixir
- #deployment
- #library
- #erlang
- #testing
- #genserver
- #mix
- #absinthe
- #remote-other
- #otp
- #plug
- #how-to-question
- #macros
- #postgres
- #channels
- #elixirconf
- #exunit
- #discussion
- #code-sync
- #javascript
- #podcasts
- #onsite
- #dialyzer
- #docker
- #authentication
- #umbrella
- #full-time-contract
- #podcasts-by-brainlid
- #ecto-query
- #elixir-ls
- #phoenix_html
- #iex
- #blog-post
- #graphql
- #genstage
- #ai
- #websockets
- #supervisor
- #advent-of-code
- #elixirconf-us
- #distillery
- #processes
- #forms
- #api
- #metaprogramming
- #security
- #performance








