I've written a routine allowing the client user to declare where their .csv import file resides and then send it to the server to import the many rows into an existing model table, and then throw it away. I've imported many graphic files into my application using attachment_fu and this was very straightforward as they were imported into an existing model object. Instead, I am trying to import a file, that when it's processed will be 'thrown away'. These examples are running ok, but I won't be able to use this strategy in my production environment... Here's my view code:
<%= form_tag(:url => {:action=> "post" }, :html => { :multipart => true }) -%> <p><%= file_field_tag "file", :size => 75, :class => "csv-input" -%></
<p><strong><label for="file">File to Upload</label></strong></p> <p><%= submit_tag -%></p>
Here's my controller code:
class ImportController < ApplicationController def aaccount_import @owner = Owner.find(2) n = 0 unless params[:file].nil? fullpath = File.expand_path(params[:file]) # this doesn't blow up but some books say that a file object is passed and thus File methods should apply? FasterCSV.foreach("#{RAILS_ROOT}/convert/#{params[:file]}") do | row> # FasterCSV.foreach(params[:file]) do |row| # cant seem to read submitting machines full path info? # FasterCSV.foreach(fullpath) do |row| c = @owner.accounts.new c.code = row[0] c.description = row[1] if c.save n += 1 else n += 1 d = Badboy.new d.message = 'Aaccount Record Failed to convert' d.recno = n d.save end end end end end
You'll notice that I am able to import this file on MY OWN machine but I'm having to HARD CODE the file path before the actual file name that is delivered. You can see some ideas I've tried to make the params[:file] behave like a FILE object and thus reveal its full path. I am grateful for any links or suggestions. Kathleen
