array structure bug/issue w/XML_Unserializer

From: erik@pixiedustgifts.com Subject: array structure bug/issue w/XML_Unserializer Date: October 30, 2004 6:55:19 PM EDT To: pear-general@lists.php.net I'm using XML_Serializer to unserialize XML from the United States Post Office APIs for shipping rates. Here's a sample XML response when trying to get a rate for one package. <RateResponse> <Package ID="myself"> <Service>Priority</Service> <ZipOrigination>20770</ZipOrigination> <ZipDestination>09021</ZipDestination> <Pounds>5</Pounds> <Ounces>1</Ounces> <Container>None</Container> <Size>REGULAR</Size> <Machinable>FALSE</Machinable> <Zone>3</Zone> <Postage>7.90</Postage> </Package> </RateResponse> And here is what you get when you unserialize it. array ( 'Package' => array (

    'ID' => 'myself',
    'Service' => 'First Class',
    'ZipOrigination' => '29487',
    'ZipDestination' => '29412',
    'Pounds' => '0',
    'Ounces' => '1',
    'Container' => 'NONE',
    'Size' => 'REGULAR',
    'Machinable' => 'FALSE',
    'Zone' => '1',
    'Postage' => '0.37',

), ) All fine and dandy, I can work with that. Now, here's the response when trying to get the rates for the same package but shipped 2 different ways. <RateResponse> <Package ID="myself"> <Service>First Class</Service> <ZipOrigination>29487</ZipOrigination> <ZipDestination>29487</ZipDestination> <Pounds>0</Pounds> <Ounces>1</Ounces> <Container>NONE</Container> <Size>REGULAR</Size> <Machinable>FALSE</Machinable> <Zone>1</Zone> <Postage>0.37</Postage> </Package> <Package ID="myself"> <Service>EXPRESS</Service> <ZipOrigination>29487</ZipOrigination> <ZipDestination>29487</ZipDestination> <Pounds>0</Pounds> <Ounces>1</Ounces> <Container>NONE</Container> <Size>REGULAR</Size> <Machinable>FALSE</Machinable> <Zone>1</Zone> <Postage>13.65</Postage> </Package> </RateResponse> And here is what you get when you unserialize it. array ( 'Package' => array (

    0 =>
    array (
      'ID' => 'myself',
      'Service' => 'First Class',
      'ZipOrigination' => '29487',
      'ZipDestination' => '29412',
      'Pounds' => '0',
      'Ounces' => '1',
      'Container' => 'NONE',
      'Size' => 'REGULAR',
      'Machinable' => 'FALSE',
      'Zone' => '1',
      'Postage' => '0.37',
    ),
    1 =>
    array (
      'ID' => 'myself',
      'Service' => 'EXPRESS',
      'ZipOrigination' => '29487',
      'ZipDestination' => '29412',
      'Pounds' => '0',
      'Ounces' => '1',
      'Container' => 'NONE',
      'Size' => 'REGULAR',
      'Machinable' => 'FALSE',
      'Zone' => '1',
      'Postage' => '13.65',
    ),

), ) See the difference? Notice how the second version now contains numerically indexed arrays while the first version did not? In short, the array structure is different for one <Package> tag than it is for two <Package> tags. I don't know if it is intended to work that way or not but I found it quite surprising and nearly impossible to loop through to get the data out if the array structure varies. I noticed someone else posted this issue and the suggestion was to use an "if/else" construct. A workable solution to me would be to make everything like the second example and have indexed arrays no matter the number of <Package> tags so for one package the result would be: array ( 'Package' => array (

    0 =>
    array (
      'ID' => 'myself',
      'Service' => 'First Class',
      'ZipOrigination' => '29487',
      'ZipDestination' => '29412',
      'Pounds' => '0',
      'Ounces' => '1',
      'Container' => 'NONE',
      'Size' => 'REGULAR',
      'Machinable' => 'FALSE',
      'Zone' => '1',
      'Postage' => '0.37',
    ),

), ) and the structure would be consistent every time. - Erik Heerlein