A platform for community discussion. Free, open, simple.
# frozen_string_literal: true
module DiscourseWorkflows
class DataTable < ActiveRecord::Base
self.table_name = "discourse_workflows_data_tables"
VALID_COLUMN_TYPES = %w[string number boolean date].freeze
MAX_COLUMN_NAME_LENGTH = 63
COLUMN_NAME_FORMAT = /\A[a-zA-Z_][a-zA-Z0-9_]*\z/
validates :name,
presence: true,
uniqueness: {
case_sensitive: false,
},
length: {
maximum: 100,
},
format: {
with: /\A[a-zA-Z_][a-zA-Z0-9_ ]*\z/,
}
scope :named, ->(names) { where("LOWER(name) IN (?)", Array(names).map(&:downcase)) }
before_destroy :drop_storage_table
# Postgres can't reorder physical columns, so the display order lives in `column_order`.
# User columns sit between `id` and the timestamps; unlisted ones keep their physical order.
def columns
system, user =
DataTables::Storage
.columns(id)
.partition { |column| DataTables::Types.system_column?(column["name"]) }
id_column, *timestamps = system
user =
user.sort_by.with_index do |column, index|
[column_order.index(column["name"]) || column_order.size, index]
end
[id_column, *user, *timestamps]
end
def user_columns
columns.reject { |column| DataTables::Types.system_column?(column["name"]) }
end
class << self
def column_name(column)
column.respond_to?(:name) ? column.name.to_s : column.fetch("name").to_s
end
def column_type(column)
column.respond_to?(:column_type) ? column.column_type.to_s : column.fetch("type").to_s
end
end
private
def drop_storage_table
DiscourseWorkflows::DataTables::Facade.drop_table!(id)
end
end
end
# == Schema Information
#
# Table name: discourse_workflows_data_tables
#
# id :bigint not null, primary key
# column_order :string default([]), not null, is an Array
# name :string(100) not null
# created_at :datetime not null
# updated_at :datetime not null
# created_by_id :integer
# updated_by_id :integer
#
# Indexes
#
# idx_dwf_data_tables_on_created_by_id (created_by_id)
# idx_dwf_data_tables_on_lower_name (lower((name)::text)) UNIQUE
# idx_dwf_data_tables_on_updated_by_id (updated_by_id)
#