Posted by Roy Kaldung on 05/25/07 20:15

CK schrieb:
> Tom wrote:
>
>> "CK" <nospams@nospams.com> wrote in message
>> news:oEC5i.24262$gk7.118002@wagner.videotron.net...
>>> Roy Kaldung wrote:
>>>
>>>> CK wrote:
>>>>> Hello Folks,
>>>>> My environment is : FreeBSD 6 , apache 1.3.7 and PHP5
>>>>> I want to parse a big XML file (39MB) . In my php.ini I set
>> memory_limit
>>>>> to 64M but I still have a error message followed :
>>>>> Fatal error: Allowed memory size of 67108864 bytes exhausted (tried to
>>>>> allocate 57 bytes) ...
>>>>>
>>>>> Your helps would be appreciated .
>>>>> Thanks
>>>> Hi,
>>>>
>>>> try to parse your XML with an SAX Parser
>>>> instead a DOM parser (-> )if possible. A Dom parser builds a complete
>>>> tree of the XML file in memory, a SAX parser is 'event driven'. For
>>>> event driven parsing see http://www.php.net/manual/en/ref.xml.php
>>>>
>>>> hth,
>>>> Roy
>>> Thank you . I tried some examples from
>>> http://www.php.net/manual/en/ref.xml.php but my XML file is really too
>>> big . Do you know is some free XML parser out there on the net can
>>> display the structure of XML file ? I visit MagicParser.com but we have
>>> to buy.
>>>
>> If you can't find anything for PHP, you might check the Perl CPAN site to
>> see if someone might already have a module you could use. Depending what
>> you needed maybe "XML" or "XML::Simple".
>>
>> Since you're running FreeBSD, it might be worth chekcing ports too.
>>
>> Tom
>
> Thanks Tom,
>
> I tried simplexml_load_file from http://ca.php.net/simplexml_load_file .
> It seems to take time to parse my XML file.
> Let me explain to you my problem :
> I have to download to my serveur every day a big XML file (about 40MB ).
> The content of this XML file is about the location,products, description,
> price, etc. . I have to write a PHP program to do the search (with
> selection criteria) about the products into this file .
> The time to parse this XML file is about 15 secs . And I do not want that
> ervery one who visit my page would wait at least 15 secs . So that I want
> is to load only once this file into an objet according to
> simplexml_load_file ( $xml_product
> =simplexml_load_file('myXMLfile.xml') ) .
> Do you know how to define that objet globally so that every one who visit my
> page does not have to read or load the XML file ?

You should try do put the parsing task into a cron job right after the
download of the file and putting the relevant contents into a database
for the search. It's probably better to spend one time 15 secs or more
than for each single request.

Roy

[Back to original message]