Ensure Rails database record uniqueness when updating without aborting the update process

Question

Ruby 2.3.0, Rails 4.2.4, and actually using postgreSQL rather than SQLite

Updated for clarity

I have a large csv file (externally updated & downloaded daily) and wrote a method to update a Rails database table. I do not want the method to append all rows to the database without validating uniqueness, so I use this great solution (How do I make a column unique and index it in a Ruby on Rails migration?) with add_index.

I'm using a rake file to store the executable update code and I enter $ rake update_task in my terminal (which works IF the table has no duplicates with the imported csv rows). The problem with this is that the database ABORTS (rake aborted!) the rake when it encounters the first duplicate entry (ERROR: duplicate key value violates unique constraint).

What can I do to remove/not save any duplicates while avoiding aborting/failing? I cannot simply drop the database table and reload it each day. Here is the schema:

ActiveRecord::Schema.define(version: 20160117172450) do

# These are extensions that must be enabled in order to support this database
enable_extension "plpgsql"

  create_table "tablename", force: :cascade do |t|
    t.string   "attr1"
    t.string   "attr2"
    t.string   "attr3"
    t.datetime "created_at", null: false
    t.datetime "updated_at", null: false
  end

  add_index "tablename", ["attr1", "attr2", "attr3"], name: "index_tablename_on_attr1_and_attr2_and_attr3", unique: true, using: :btree

end

and my rake task in lib/tasks/file_name.rake contents:

desc "Download data and update database table"

task :update_task => :environment do
  u = CorrectClassName.new
  u.perform_this
end

and CorrectClassName is in an .rb file in app/directory1:

class CorrectClassName

  def perform_this
    something = ClassWithUpdateCode.new
    something.update_database
  end

end

and ClassWithUpdateCode is in an .rb file in app/directory2:

require 'csv'

class ClassWithUpdateCode

  def update_database
    csv_update = File.read(Rails.root.join('lib', 'assets', "file_name.csv"))
    options = {:headers => true}

    csv = CSV.parse(csv_update, options)
    csv.each do |row|
        tm = TableModel.new

        tm.attr1 = row[0]
        tm.attr2 = row[1]
        tm.attr3 = row[2]
        tm.save # maybe I can use a different method or if statement here?
    end
  end

end

Update: @Kristan's solution works below, but here is where to put the begin/rescue/end handling:

In the .rb file in app/directory2:

require 'csv'

class ClassWithUpdateCode

  def update_database
    csv_update = File.read(Rails.root.join('lib', 'assets', "file_name.csv"))
    options = {:headers => true}

    csv = CSV.parse(csv_update, options)
    csv.each do |row|
        tm = TableModel.new
        begin
          tm.attr1 = row[0]
          tm.attr2 = row[1]
          tm.attr3 = row[2]
          tm.save
        rescue ActiveRecord::RecordNotUnique
        end
    end
  end

end

Answer 1

rake is bailing out because an exception gets raised when you try to save a record that violates your table's uniqueness constraint. The simplest way to prevent that is by catching and ignoring the exception. I'm assuming your record is created during u.perform_this.

task :update_task => :environment do
  u = CorrectClassName.new
  begin
    u.perform_this
  rescue ActiveRecord::RecordNotUnique
    # move on
  end
end

Another option is to add a uniqueness validation to your Rails model, then either check valid? before saving or call create (not create!), which doesn't raise validation exceptions.

class CorrectClassName < ActiveRecord::Base
  validates_uniqueness_of :attr1, scope: [:attr2, :attr3]
end

task :update_task => :environment do
  u = CorrectClassName.new(data)
  u.perform_this if u.valid?
end