FasterCSV: Check whether a file is invalid before accepting it - is there a simpler way?

Question

I'm using FasterCSV on a Ruby on Rails application and currently it throws an Exception if the file is invalid.

I've looked over the FasterCSV doc, and it seems that if I use FasterCSV::parse with a block, it'll read the file one line at a time, without allocating too much memory. It'll throw a FasterCSV::MalformedCSV exception if there is any kind of error on the file.

I've implemented a custom solution, but I'm not sure it's the best possible one (see my answer below). I'd be interested in knowing alternatives

Answer 1

I made some tests yesterday and it turns out that my solution didn't quite work; I kept getting empty arrays on valid CSVs after implementing the first is_valid . I'm not sure whether it's a FasterCSV caching issue or something in my code, and I don't know if it's related with my test setup, but I decided to go implement a safe_parse instead:

#/lib/faster_csv_safe_parse.rb
class FasterCSV

  def self.safe_parse(file, options = {})
    begin
      FasterCSV.parse(file, options)
    rescue FasterCSV::MalformedCSVError
      nil
    end
  end

end

This will return a parsed array if the file is valid, or nil otherwise. I could then implement my validations as follows:

# /models/csv_importer.rb

class CsvImporter
  include ActiveRecord::Validations

  validates_presence_of :file
  validate check_file_format
  attr_accessor csv_data

  def csv_data
    @csv_data ||= FasterCSV.safe_parse(file)
  end

...

  private

  def check_file_format
    errors.add :file, "Malformed CSV! Please check syntax" if csv_data.nil?
  end
end

I guess it would be possible to implement a safe_parse that accepts a block and parses the file line by line, but for my purposes this simple implementation was enough, and it works in all cases.