CSVImporter

Importing a CSV file is easy to code until real users attempt to import real data.

CSVImporter aims to handle validations, column mapping, import and reporting.

Rationale

Importing CSV files seems easy until you deal with real users uploading their real CSV file. You then have to deal with ASCII-8BIT formats, missing columns, empty rows, malformed headers, wild separators, etc. Reporting progress and errors to the end-user is also key for a good experience.

I went through this many times so I decided to build CSV Importer to save us the trouble.

CSV Importer provides:

a DSL to define the mapping between CSV columns and your model
good reporting to the end user
support for wild encodings and CSV formats.

It is compatible with ActiveRecord 4+ and any ORM that implements the class methods transaction and find_by and the instance method save.

Usage tldr;

Define your CSVImporter:

class ImportUserCSV
  include CSVImporter

  model User # an active record like model

  column :email, to: ->(email) { email.downcase }, required: true
  column :first_name, as: [ /first.?name/i, /pr(é|e)nom/i ]
  column :last_name,  as: [ /last.?name/i, "nom" ]
  column :published,  to: ->(published, user) { user.published_at = published ? Time.now : nil }

  identifier :email # will update_or_create via :email

  when_invalid :skip # or :abort
end

Run the import:

import = ImportUserCSV.new(file: my_file)

import.valid_header?  # => false
import.report.message # => "The following columns are required: email"

# Assuming the header was valid, let's run the import!

import.run!
import.report.success? # => true
import.report.message  # => "Import completed. 4 created, 2 updated, 1 failed to update"

Installation

Add this line to your application's Gemfile:

gem 'csv-importer'

And then execute:

$ bundle

Or install it yourself as:

$ gem install csv-importer

Usage

Create an Importer

Create a class and include CSVImporter.

class ImportUserCSV
  include CSVImporter
end

Associate an active record model

The model is likely to be an active record model.

class ImportUserCSV
  include CSVImporter

  model User
end

It can also be a relation which is handy to preset attributes:

class User
  scope :pending, -> { where(status: 'pending') }
end

class ImportUserCSV
  include CSVImporter

  model User.pending
end

You can change the configuration at runtime to scope down to associated records.

class Team
  has_many :users
end

team = Team.find(1)

ImportUserCSV.new(path: "tmp/my_file.csv") do
  model team.users
end

Define columns and their mapping

This is where the fun begins.

class ImportUserCSV
  include CSVImporter

  model User

  column :email
end

This will map the column named email to the email attribute. By default, we downcase and strip the columns so it will work with a column spelled " EMail ".

Now, email could also be spelled "e-mail", or "mail", or even "courriel" (oh, canada). Let's give it a couple of aliases then:

  column :email, as: [/e.?mail/i, "courriel"]

Nice, emails should be downcased though, so let's do this.

  column :email, as: [/e.?mail/i, "courriel"], to: ->(email) { email.downcase }

If you need to do more advanced stuff, you've got access to the model:

  column :email, as: [/e.?mail/i, "courriel"], to: ->(email, user) { user.email = email.downcase; model.super_user! if email[/@brewhouse.io\z/] }

Like very advanced stuff? We grant you access to the column object itself which contains the column name – quite handy if you want to support arbitrary columns.

  column :extra, as: [/extra/], to: ->(value, user, column) do
    attribute = column.name.sub(/^extra /, '')
    user[attribute] = value
  end

Note that to: accepts anything that responds to call and take 1, 2 or 3 arguments.

class ImportUserCSV
  include CSVImporter

  model User

  column :birth_date, to: DateTransformer
  column :renewal_date, to: DateTransformer
  column :next_renewal_at, to: ->(value) { Time.at(value.to_i) }
end

class DateTransformer
  def self.call(date)
    Date.strptime(date, '%m/%d/%y')
  end
end

Now, what if the user does not provide the email column? It's not worth running the import, we should just reject the CSV file right away. That's easy:

class ImportUserCSV
  include CSVImporter

  model User

  column :email, required: true
end

import = ImportUserCSV.new(content: "name\nbob")
import.valid_header? # => false
import.report.status # => :invalid_header
import.report.message # => "The following columns are required: 'email'"

Update or Create

You often want to find-and-update-or-create when importing a CSV file. Just provide an identifier, and we'll do the hard work for you.

class ImportUserCSV
  include CSVImporter

  model User

  column :email, to: ->(email) { email.downcase }

  identifier :email
end

And yes, we'll look for an existing record using the downcased email. :)

You can also define a composite identifier:

  # Update records with matching company_id AND employee_id
  identifier :company_id, :employee_id

Or a Proc:

  # Update records with email if email is present
  # Update records matching company_id AND employee_id if email is not present
  identifier ->(user) { user.email.empty? ? [:company_id, :employee_id] : :email }

Skip or Abort on error

By default, we skip invalid records and report errors back to the user. There are times where you want your import to be an all or nothing. The when_invalid option is here for you.

class ImportUserCSV
  include CSVImporter

  model User

  column :email, to: ->(email) { email.downcase }

  when_invalid :abort
end

import = ImportUserCSV.new(content: "email\nbob@example.com\nINVALID_EMAIL")
import.valid_header? # => true
import.run!
import.report.success? # => false
import.report.status # => :aborted
import.report.message # => "Import aborted"

You are now done defining your importer, let's run it!

Import from a file, path or string

You can import from a file, path or just the CSV content. Please note that we currently load the entire file in memory. Feel free to contribute if you need to support CSV files with millions of lines! :)

import = ImportUserCSV.new(file: my_file)
import = ImportUserCSV.new(path: "tmp/new_users.csv")
import = ImportUserCSV.new(content: "email,name\nbob@example.com,bob")

Overwrite configuration at runtime

It is often needed to change the configuration at runtime, that's quite easy:

team = Team.find(1)
import = ImportUserCSV.new(file: my_file) do
  model team.users
end

`after_build` and `after_save` callbacks

You can preset attributes (or perform any changes to the model) at configuration or runtime using after_build

class ImportUserCSV
  model User

  column :email

  after_build do |user|
    user.name = email.split('@').first
  end
end

# assuming `current_user` is available

import = ImportUserCSV.new(file: my_file) do
  after_build do |user|
    user.created_by_user = current_user
  end
end

The after_save callback is run after each call to the method save no matter it fails or succeeds. It is quite handy to keep track of progress.

progress_bar = ProgressBar.new

UserImport.new(file: my_file) do
  after_save do |user|
    progress_bar.increment
  end
end

Skip import

You can skip the import of a model by calling skip! in an after_build block:

UserImport.new(file: csv_file) do
  # Skip existing records
  after_build do |user|
    skip! if user.persisted?
  end
end

Validate the header

On a web application, as soon as a CSV file is uploaded, you can check if it has the required columns. This is handy to fail early and provide the user with a meaningful error message right away.

import = ImportUserCSV.new(file: params[:csv_file])
import.valid_header? # => false
import.report.message # => "The following columns are required: "email""

Run the import and provide feedback to the user

import = ImportUserCSV.new(file: params[:csv_file])
import.run!
import.report.message  # => "Import completed. 4 created, 2 updated, 1 failed to update"

You can get your hands dirty and fetch the errored rows and the associated error message:

import.report.invalid_rows.map { |row| [row.line_number, row.model.email, row.errors] }
  # => [ [2, "INVALID_EMAIL", { "email" => "is invalid" } ] ]

We do our best to map the errors back to the original column name. So with the following definition:

  column :email, as: /e.?mail/i

and csv:

E-Mail,name
INVALID_EMAIL,bob

The error returned should be: { "E-Mail" => "is invalid" }

Custom quote char

You can handle exotic quote chars with the quote_char option.

email,name
bob@example.com,'bob "elvis" wilson'

import = ImportUserCSV.new(content: csv_content)
import.run!
import.report.status
  # => :invalid_csv_file
import.report.messages
  # => CSV::MalformedCSVError: Illegal quoting in line 2.

Let's provide a valid quote char:

import = ImportUserCSV.new(content: csv_content, quote_char: "'")
import.run!
  # => [ ["bob@example.com", "bob \"elvis\" wilson"] ]

Custom encoding

You can handle exotic encodings with the encoding option.

ImportUserCSV.new(content: "メール,氏名".encode('SJIS'), encoding: 'SJIS:UTF-8')

Development

After checking out the repo, run bin/setup to install dependencies. Then, run bin/console for an interactive prompt that will allow you to experiment.

To install this gem onto your local machine, run bundle exec rake install. To release a new version, update the version number in version.rb, and then run bundle exec rake release to create a git tag for the version, push git commits and tags, and push the .gem file to rubygems.org.

Contributing

Fork it ( https://github.com/pcreux/csv-importer/fork )
Create your feature branch (git checkout -b my-new-feature)
Commit your changes (git commit -am 'Add some feature')
Push to the branch (git push origin my-new-feature)
Create a new Pull Request

pcreux/csv-importer