使用CSV数据更新XML的Perl脚本报错及功能实现问题

Let's break down and fix the issues you're facing:

Key Problems in the Original Script #

• The name value isn't a separate tag—it's an attribute of the root <ns0:networkdevice> element, so look_down('_tag' => 'name') returns undef, causing the push_content error.
• Your code was trying to output the original unmodified XML tree ($tree) instead of each processed template for every CSV row.
• The ipaddress lookup wasn't targeting the correct nested element (it's buried inside <NetworkDeviceIP> under <NetworkDeviceIPList>).

Modified Working Script #

#!/usr/bin/perl
use warnings;
use strict;
use Text::CSV;
use XML::TreeBuilder;

my $xmlbase = 'template.xml';
my $paramsfile = 'input.csv';
# Map CSV columns to their XML target: attribute or nested tag
my @field_map = (
    { type => 'attr', target_tag => 'networkdevice', attr_name => 'name' },
    { type => 'tag',  target_tag => 'ipaddress', parent_path => ['NetworkDeviceIPList', 'NetworkDeviceIP'] }
);

my $csv = Text::CSV->new({ binary => 1 }) or die "Cannot use CSV: " . Text::CSV->error_diag();
open my $fh, "<:encoding(utf8)", $paramsfile or die "$paramsfile: $!";

# Uncomment below if your Input.csv has a header row (e.g., Name, IP Address)
# $csv->getline($fh);

while ( my $row = $csv->getline($fh) ) {
    # Load fresh template for each device to avoid cross-row contamination
    my $contree = XML::TreeBuilder->new();
    $contree->parsefile($xmlbase);
    
    my $index = 0;
    foreach my $field (@field_map) {
        my $value = $row->[$index++];
        next unless defined $value;
        
        if ($field->{type} eq 'attr') {
            # Update root element's name attribute
            my $root = $contree->look_down('_tag' => $field->{target_tag});
            $root->attr($field->{attr_name}, $value) if $root;
        } elsif ($field->{type} eq 'tag') {
            # Traverse nested path to find the parent of ipaddress
            my $parent_node = $contree;
            foreach my $path_part (@{$field->{parent_path}}) {
                $parent_node = $parent_node->look_down('_tag' => $path_part);
                last unless $parent_node;
            }
            # Replace existing ipaddress content with CSV value
            if ($parent_node) {
                my $ip_tag = $parent_node->look_down('_tag' => $field->{target_tag});
                $ip_tag->replace_content($value) if $ip_tag;
            }
        }
    }
    
    # Output the processed XML for this device (or save to a file)
    print '<?xml version="1.0" encoding="UTF-8"?>' . "\n";
    print $contree->as_XML() . "\n\n";
    
    # Clean up to prevent memory leaks
    $contree->delete();
}

close $fh;

What Changed & Why #

1. Field Mapping Structure: We use an array of hashes to clearly define if each CSV field targets an XML attribute or a nested tag. This makes the script easier to adjust if your XML structure changes later.
2. Handling the name Attribute: We directly target the root <ns0:networkdevice> element (XML::TreeBuilder automatically strips namespace prefixes from tag names) and use the attr() method to set its name property—this fixes the push_content error entirely.
3. Nested ipaddress Update: We traverse the XML hierarchy step-by-step (NetworkDeviceIPList → NetworkDeviceIP) to find the correct parent of the ipaddress tag, then use replace_content() to overwrite its default value (instead of appending to it with push_content).
4. Per-Row Template Freshness: We load a new template instance for every CSV row to ensure changes from one device don't leak into the next.
5. Memory Management: Added $contree->delete() after processing each device to free up memory, which is helpful if you have a large CSV file.

Optional Enhancements #

• If your Input.csv includes a header row, uncomment the line that skips it to avoid processing the header as a device entry.
• Instead of printing to the console, you could save each processed XML to a unique file (e.g., LAB-1.xml) by adding file-handling code inside the loop.

内容的提问来源于stack exchange，提问作者StayCalm