EOF pattern matching

[mThomas]

I have an XML file which appears to be generated from a Pivot Table. What I need to do is find an instance of something in the each <z:row> element. If the something exists, remove and push the element to a separate file. After each element with the something in it has been removed from the original file, the original file is sent to a different department and my work with that portion of the file is done.

Here is a snippet of the file:

<rs:data>
<z:row City='Eastover' State='NC' Due_Date='2005-10-11T00:00:00'
		 Due_Time='1899-12-30T15:00:00' Name='Eastover Central Recreation center'
		 Description='Construction of the Eastover Central Recreation Center Project'
		 Estimated_Cost='0' Pre_Bid='True' Mandatory='False' Pre_Bid_Date='2005-09-29T00:00:00'
		 Pre_Bid_Time='1899-12-30T15:00:00' Building='False' Region='0'
		 Owner='0' Architect='0' Engineer='0' Other='0' Specs='True'
		 Plans='True' Processed='1899-12-30T00:00:00' Firm='MVcdc'
		 OtherProjectNumber='' Other1='' Other2='' CDC='True' CDCNumber='NC Q000000001'
		 Job_Key='' CDCURL='[URL unfurl="true"]http://www.aeplans.com/aeplans/cdc/transfer.asp?ID=LQD3492;27/OXefe'[/URL]
		 CDCVERIFYURL='[URL unfurl="true"]http://www.aeplans.com/aeplans/cdc/verify.asp?ID=LQD3492;27/OXefe'/>[/URL]

There are multiple such row elements. I can pattern match for the “something”. The something is CDC=’False’. If that occurs then I want to remove the element from the xml file.

I was thinking one way would be to loop through the file and push everything between a < and a > into a variable. I would then check if the something existed. If the something exists then print the variable to a new file. If the something does not exist, then print the variable to a different file. I would end up with two separate files. One file with all the row elements without the something and a file with all the row elements with the something.

I am trying to use a regular expression to match all occurrences of any text/characters/numbers between the < and >. Here is part of the code I’m trying. This is just trying to see if I can grab all of the code between each < and > in the file.

$xml_file="cdc10-31-2005-8-56-20.xml";
open(XML, $xml_file) || die("Could not open file!");
@xml=<XML>;
close(XML);

foreach (@xml) {

$file = m/<(.+)>/;

print "$1\n";

}

This only returns rs:data for each line. That in and of its self confused me as there is only one <rs:data> and one </rs:data> in the file. I am presuming this is a root tag.

Is there a way to tell my pattern match to match until it gets to a > even if there are new lines and tabs? This part of my code $file = m/<(.+)>/; does not do that.

I hope I have given enough information.

As always, any help will be greatly appreciated.

mike

 

[PaulTEG]

Have a look on

http://search.cpan.org,

there's a plethora of XML parsers out there

HTH
--Paul

Spend an hour a week on CPAN, helps cure all known programming ailments ;-)

 

[mThomas]

I have a couple of xml IDE's. I was hoping to not have to use one and do a simple pattern matching.

My pattern matching question is as much about future reference as it is about this project.

tia,

mmike

 

[KevinADC]

this shouldn't be too difficult if the element were all on one line, will it matter if the <z:row> tag is transformed into a single line instead of broken over several lines? Then you could do seomthing like:

if (/^<z:row/ and /CDC=’False’/ and /\/>/) {
   do something
}

 

[dmazzini]

YOu could use the module XML::Simple

There are good articles...Type in google

Perl and XML


dmazzini
GSM System and Telecomm Consultant

 

[mThomas]

I tried using XML::Simple and it would probably work, however considering I don't really have to read the file as XML, here is what I did.

# more code here

$current = 0;
$next = 1;
$variable = "";

foreach(@xml){
	
	if($_ =~ /<z:row/) { 

		if($next != $current) { $variable = $_; $next = $current;}

	} # end if z row
				
	if($current = $next) { $variable = $variable.$_;  }

	if($_ =~ />/) {

		 push @onearray, $variable;
		$variable = "";
		$next = 1;

	} # end if />

}

foreach(@onearray) {

if($_ =~ /CDC='False'/) {push @false, $_;}
if($_ =~ /CDC='True'/) {push @true, $_;}

}

# more code here

It's not sweet and there is probably a regex that would have done a sweeter job, but it works.

mike

 

[KevinADC]

that looks OK. You could maybe shorten the last foreach loop up a bit if those are the only two conditions, you really only need check for one condition when there are only two possible (boolean):

(/CDC='False'/ ? push @false,$_ : push @true,$_) for (@onearray);

the terneary operator is perfect for boolean conditions:

if this part is true ? do this : else do this