Complex XML parsing with Perl and LIBXML - perl

Question

I have XML:
<?xml version="1.0" encoding="UTF-8"?>
<?xml-stylesheet type="text/xsl" href="MeasDataCollection.xsl"?>
<measCollecFile xmlns="">
<fileHeader fileFormatVersion="32.435 V7.2.0">
</fileHeader>
<measData>
<managedElement localDn="bs=8" swVersion="R21A"/>
<measInfo measInfoId="CORE,SIP_session_statistics">
<measType p="1">CPUUSAGE</measType>
<measType p="2">CPUMEM</measType>
<measType p="3">SYSMEM</measType>
<measValue measObjLdn="SGC.bsNo=17,networkRole=2">
<r p="1">10</r>
<r p="2">20</r>
<r p="3">30</r>
</measValue>
<measValue measObjLdn="SGC.bsNo=18,networkRole=2">
<r p="1">40</r>
<r p="2">50</r>
<r p="3">60</r>
</measValue>
</measInfo>
</measData>
</measCollecFile>
QUESTION:
I want to extract the 40 from <r p="1">40</r> element. The only thing given is <measType p="1">CPUUSAGE</measType> and <measValue measObjLdn="SGC.bsNo=18,networkRole=2">
i.e. I only know that I need to find the CPUUSAGE of the bsNo=18. The order of the data is always maintained.
Here is what I have tried so far:
my $qry="//measInfo[measType/text() = 'CPUUSAGE']/measValue";
my #nodes= $conn->findnodes($qry);
foreach my $vnode (#nodes) {
if ($vnode->getAttribute('measObjLdn') =~ /'bsNo=18'/) {
foreach my $node ($vnode) {
foreach my $p ($node->getChildnodes) {
if (ref($p)=~'Element'){
$no=$p->textContent;
print $no;**#this prints the value of all the <r> elements**
}
}
}
}
}
My challenge is there can be many elements like CPUUSAGE,CPUMEM... and how I can reach the correct order in the <r> element in that order for a given measValue attribute (/'bsNo=18'/).
And subsequently modify that 40 to some other desired value**

Answer 1

Your Perl code can't work because you match the attribute value against 'bsNo=18' including single quotes.
If you want to find the r element with the same p attribute as the CPUUSAGE node, you could either try the single XPath expression by ikegami or something like the following:
for my $type_node ($conn->findnodes('//measInfo/measType[.="CPUUSAGE"]')) {
my $p = $type_node->getAttribute('p');
my $qry = <<"EOF";
..
/measValue[contains(concat(\#measObjLdn, ','), 'bsNo=18,')]
/r[\#p='$p']
EOF
for my $r_node ($type_node->findnodes($qry)) {
print $r_node->textContent, "\n";
}
}
This first loops over all measType nodes whose content is CPUUSAGE, gets the p attribute then finds all the corresponding r nodes. This approach should be more efficient than a single XPath query.
To find the r node by position and modify its contents, try:
for my $type_node ($conn->findnodes('//measInfo/measType[.="CPUUSAGE"]')) {
my $pos = $type_node->findvalue('count(preceding-sibling::measType) + 1');
my $qry = <<"EOF";
..
/measValue[contains(concat(\#measObjLdn, ','), 'bsNo=18,')]
/r[$pos]
EOF
for my $r_node ($type_node->findnodes($qry)) {
$r_node->removeChildNodes;
$r_node->appendText('50');
}
}
print $conn->toString;