XML import Question

burguer · March 15, 2010, 10:02am

hi guys can somebody help to find some tutorial or link to find the way
to import an external XML file to rails with some nodes and child-nodes.

Thanks

burguer · March 15, 2010, 2:55pm

Try using Nokogiri: http://nokogiri.org/ It’s a great and easy to use
parser.

burguer · March 16, 2010, 12:38am

What are you after? The answer really depends on your situation.

Two options come to mind:

A custom rake task. You can see a simple example here that uses YAML
instead of XML as source data:

http://github.com/kete/kete/blob/master/lib/tasks/user_export_import.rake

Here’s an example of parsing from source XML (modify this and replace
the YAML parsing the above rake task):

github.com

kete/kete/blob/master/lib/workers/excel_based_importer_worker.rb

require 'importer'

require 'nokogiri'
require 'logger'
class ExcelPreProcessor < Nokogiri::XML::SAX::Document
  def initialize(path_to_xl_xml_file, options = {})
    # maybe TODO: refactor assignment from options
    # this fixes cut and paste error when code was moved to separate class
    # but ExcelPreProcessor may know too much about the calling environment
    @zoom_class = options[:zoom_class]
    @import_dir_path = options[:import_dir_path]
    @record_interval = options[:record_interval]

    @logger = Logger.new('/home/kete/excel-parse.log')
    excel_xml = File.read(path_to_xl_xml_file)

    # Wrap all data in CData tags to prevent parsing issues unless they already have one
    excel_xml.gsub!(/<(?:ss\:)?Data[^>]*>(.*)<\/(?:ss\:)?Data>/i) do |match|
      value = $1
      value = "<![CDATA[#{value.gsub('&amp;', '&')}]]>" unless value =~ /^<\!\[CDATA\[/i

This file has been truncated. show original

starting around line 30, note the hardcoding of element names that
you need to adjust

Another option, if you have already created a RESTful interface that
accepts XML for creating new records, you might simply create a script
that parses your existing XML file and submits a POST to your create
action with the relevant data. Not necessarily the fastest process,
but if it is a one off import it might be the quickest thing to put
together programming wise. Net::HTTP and Nokogiri seem like all you
really need.

I would probably just go with the rake task. Pretty simple.

Cheers,
Walter

burguer · March 16, 2010, 3:09am

Thanks a lot for your help i’ll try with thats options
have you guys ever used the "Hash.from_xml " method to import XML

thanks for the help i really really appreciate

burguer · March 16, 2010, 3:43am

On Tue, Mar 16, 2010 at 3:09 PM, Burguer B. [email protected]
wrote:

Thanks a lot for your help i’ll try with thats options
have you guys ever used the "Hash.from_xml " method to import XML

I’ve used it in the past. If you aren’t importing a huge amount of
records it might be ok. Otherwise I think you are better off using
Nokogiri. Depends on the complexity of your source XML.

Cheers,
Walter