Rails Parse CSV with empty cells and correctly handle exceptions - ruby-on-rails

Question

I am trying to allow users to upload csv files, however I am having some issues trying to figure out how to handle errors from the file itself.
My controller method:
def create
Product.import_csv(file)
redirect_to products_path, :flash => { :notice => "Items Added!" }
end
My model method to import the file:
def self.import_csv(file)
csv = CSV.read(file.path), :headers => true)
csv.each do |row|
item_id = row[0]
start_date = Date.parse(row[1])
order_date = Date.parse(row[2])
new_rec = where(item_id:item_id, order_date:order_date).first_or_initialize
new_rec.save!
end
end
All this works well when the file is properly formatted, however Im confused as to how to handle exceptions. Once such exception is when start_date or order_date are missing; I get an no implicit conversion of nil into String because I'm attempting to parse the date of an empty cell. Even though I have validations in my model for presence, they only get fired on the save action.
I don't want to silently ignore these errors with a rescue block, but instead redirect back, and notify the user.
How can I handle such exceptions, so that the jobs fails, and the user gets notified of the error, not specifically solely the error given above, but including errors we can't account for? Another example would be a empty line or something we cant even account for. How can I handle such errors, and notify the user with, for example, a generic "Bad data" message?
Should I handle this in my model or my controller?

Answer 1

If you have presence validations for the dates, you can simply assign them only if they exist, and let your model do the rest:
def self.import_csv(file)
csv = CSV.read(file.path), :headers => true)
csv.each do |row|
item_id = row[0]
start_date = row[1].nil? ? nil : Date.parse(row[1])
order_date = row[2].nil? ? nil : Date.parse(row[2])
new_rec = where(item_id:item_id, order_date:order_date).first_or_initialize
new_rec.save!
end
end