Import XML Data Topic

Monday 20 February 2006 12:41:06 am

I don't think so Guillaume because of the way data are extracted from the xml import file. Extracted data are text nodes and in case you have other xml tags inside your data they will not be considered as text nodes. Actually the parser/xpath library I use in PHP is quite poor in my opinion. We should wait for PHP 5 to rewrite this extension in a better way.

You may give a try using escaped characters but I am pessimistic.

Olivier

Friday 31 March 2006 11:33:35 am

Olivier,
I seem to have found a minor bug. There are portions of the content tree that disappear after I import a file. This is a temporary problem until I reset cache or publish a node. This is only in admin and only in the left tree view. The reason it is frustrating is that I want to go to a node following my import, but its not there! It doesn't affect the public site.

Friday 16 June 2006 10:40:48 am

Based on Samuel post I patched the extension "Import XML Data Topic" for a support of the data type 'ezxmltext'. It is a bit of a hack because I had to define a fake 'http' class in order to use ez xml validation. Here the code:

add this case statement in 'importXMLDatafunctioncollection.php'

case 'ezxmltext':
  if( $xml_content_string ) // should be outside the 'switch' statement
  {
    // Get the embedded data by cutting out everything between the escape tokens
    $begin_escape_string = '<![CDATA[';
    $end_escape_string = ']]>';
							
    $start_pos = strpos($xml_content_string, $begin_escape_string );
    $end_pos   = strpos($xml_content_string, $end_escape_string );
 							
    if( $start_pos && $end_pos )
    {
      $embedded_content = substr( $xml_content_string, $start_pos + strlen( $begin_escape_string) , ( $end_pos - $start_pos - strlen($begin_escape_string) ) ); 

      // ez allows a data validation for the datatype ezxmltext
require_once('kernel/classes/datatypes/ezxmltext/handlers/input/ezsimplifiedxmlinput.php');

      // but unfortunately the method expect a "http" object and the
      // 'contentobjectattribute'
      // so here a lot of code just to make this function happy
								
      $fakeHttp = new fakeHttp( $embedded_content );

      $myhandler = new eZSimplifiedXMLInput(); // get the handler for input validation

      $myhandler->validateInput( $fakeHttp, '', $attribute ); // tons of magic happens here
								
       // I don't have the set the value to the attribute
       // the function 'validateInput' does it already
       //$attribute->setAttribute( 'data_text', $embedded_content );
								
       // debug the $attribute object
       //echo "<pre>";
       //print_r( $attribute );
       //echo "</pre>";
    }
  }

   break;

add this class to the file

class FakeHttp
{

	var $data;

	function FakeHttp( $data )
	{
		$this->data = $data;
	}
	
	function postVariable ( $whatever )
	{
		return $this->data;
	}
	
	function hasPostVariable( $whatever )
	{
		return true;
	}
}

http://www.mugo.ca
Mugo Web, eZ Partner in Vancouver, Canada

Sunday 18 June 2006 5:41:00 am

Another thing: I can't see the sense of the $i counter. If you import many data you will fill your memory because all the XML data of all the just saved objects is still available. Just replace

foreach ($paths_item as $path_item) {
    foreach($fieldNameList as $fieldName) {
        $fieldValue = $xPathEngine->wholeText("$path_item/$fieldName"."[1]");
        $parsedItems[$i][$fieldName]=$fieldValue;
        if ($fieldName=='account-login') $isAcountObject=true;
    }

with

foreach ($paths_item as $path_item) {
    $parsedItems = array();  // resets the parsed data 
    foreach($fieldNameList as $fieldName) {
        $fieldValue = $xPathEngine->wholeText("$path_item/$fieldName"."[1]");
        $parsedItems[$fieldName]=$fieldValue; // delete the [$i] everywhere
        if ($fieldName=='account-login') $isAcountObject=true;
    }

Have a nice day and of course: thanks for the extension - saves much work :-)

Thomas

Edit: Sorry, I didn't read the next lines:

$result['data']= $parsedItems;

But I don't know if you really want that feature in a mass import situation...

Tuesday 25 July 2006 10:45:44 am

some parts of code can be taken from cronjobs/rss_import.php

{set-block scope=root variable=cache_ttl}0{/set-block}

Wednesday 06 September 2006 5:58:27 am

Hi,

You have to download and install the xpath lib (php file).

And yes it works and yes that's useful.

X+

http://www.sydesy.com

Thursday 14 September 2006 11:26:22 pm

Hi to all!
I have recently downloaded ImportXMLData.tar.
I think,I have done everything in a right way,according to installation instruction,but when I import xml_sample.xml,the data that appears above is empty.Help me,please!

Monday 30 October 2006 2:37:03 pm

Hi

I needed quite some extra features and started rebuilding some code this weekend. I ended up completely changing the php interface, the fetch call, the gui, the inifile. So eventually I renamed the whole extension to avoid confusion. I have it, local, here, it's nowhere for download at this date.

The thing still supports the original datatypes from ImportXMLData, plus a few extra, .. ezxmltext and ezobjectrelationlist amongst others. But most importantly, adding support for other datatypes is really easy now.

So how to proceed ? I can send you (Olivier) the sources so you can look at it and see if you want to pick it up ? I can't find your email address or PM you in this forum, so you may contact me at pike-ez [AT] kw.nl, if you like ?

Alternatively, I could post it as a new contribution. But having 2 xmlimporters is confusing too.

there are some screenshots here
http://pike.kw.nl/browse/files/projects/pike/2006/ezpublish/XMLImport/screensnaps

$2c,
*pike

---------------
The class eZContentObjectTreeNode does.

Thursday 02 November 2006 2:33:28 pm

Hi

Esu .. interesting .. how do you import images ? the version i had only imported numbers,strings,boolean,dates,..and user. i need support for files and images myself .. planned to write it.

& I'm curious: how do you recognize if one object was already imported ? do you keep a node_id or object_id in your xml, or some ini setting for a unique id ?

curiousss,
*-pike

---------------
The class eZContentObjectTreeNode does.

Thursday 02 November 2006 11:59:21 pm

Hi

Here's how we handle import of eZXMLType datatype in the RSS import.

function setEZXMLAttribute( &$attribute, &$attributeValue, $link = false )
{
    include_once( 'kernel/classes/datatypes/ezxmltext/handlers/input/ezsimplifiedxmlinputparser.php' );
    $contentObjectID = $attribute->attribute( "contentobject_id" );
    $parser = new eZSimplifiedXMLInputParser( $contentObjectID, false, 0 );

    $attributeValue = str_replace( "\r", '', $attributeValue );
    $attributeValue = str_replace( "\n", '', $attributeValue );
    $attributeValue = str_replace( "\t", ' ', $attributeValue );

    $document = $parser->process( $attributeValue );
    if ( !is_object( $document ) )
    {
        $cli =& eZCLI::instance();
        $cli->output( 'Error in xml parsing' );
        return;
    }
    $domString = eZXMLTextType::domString( $document );

    $attribute->setAttribute( 'data_text', $domString );
    $attribute->store();
}

See also : <i>cronjobs/rssimport.php:490</i>

Kåre Høvik