XML-zu-CSV-Konverter

// Browser or Deno. DOMParser never resolves external entities, so XXE is
// not reachable here. In Node use @xmldom/xmldom, which also does not.
function xmlToCsv(source, rowTag) {
  const doc = new DOMParser().parseFromString(source, 'application/xml');
  const error = doc.querySelector('parsererror');
  if (error) throw new Error(error.textContent.trim());

  const root = doc.documentElement;
  const rows = [...root.children].filter((el) => el.tagName === rowTag);
  if (rows.length === 0) throw new Error('No <' + rowTag + '> under <' + root.tagName + '>');

  const columns = [];
  const records = rows.map((row) => {
    const record = {};
    const put = (key, value) => {
      if (!columns.includes(key)) columns.push(key);
      // Repeated leaves are joined. That is lossy; count it in real code.
      record[key] = record[key] === undefined ? value : record[key] + '|' + value;
    };
    const walk = (el, prefix) => {
      for (const a of el.attributes) {
        put(prefix ? prefix + '.@' + a.name : '@' + a.name, a.value);
      }
      const kids = [...el.children];
      if (kids.length === 0) {
        put(prefix || el.tagName, el.textContent.trim());
        return;
      }
      for (const k of kids) walk(k, prefix ? prefix + '.' + k.tagName : k.tagName);
    };
    walk(row, '');
    return record;
  });

  // RFC 4180: quote a field containing a comma, a quote or a line break,
  // and double any quote inside it.
  const cell = (v) => (/[",\r\n]/.test(v) ? '"' + v.replace(/"/g, '""') + '"' : v);
  const line = (values) => values.map((v) => cell(v ?? '')).join(',');

  return [line(columns), ...records.map((r) => line(columns.map((c) => r[c])))].join('\r\n');
}

# pip install defusedxml
# The standard library parser is not safe against entity expansion.
# defusedxml is a drop-in replacement that closes that and external entities.
import csv
import sys
from defusedxml.ElementTree import parse


def flatten(el, prefix, record, columns):
    for name, value in el.attrib.items():
        key = f"{prefix}.@{name}" if prefix else f"@{name}"
        if key not in columns:
            columns.append(key)
        record[key] = value

    children = list(el)
    if not children:
        key = prefix or el.tag
        if key not in columns:
            columns.append(key)
        record[key] = (el.text or "").strip()
        return

    for child in children:
        # ElementTree reports namespaced names in Clark notation,
        # '{http://purl.org/dc/elements/1.1/}title', not 'dc:title'.
        tag = child.tag.split("}")[-1]
        key = f"{prefix}.{tag}" if prefix else tag
        if key in record:  # repeated leaf: join, and note the loss
            record[key] += "|" + (child.text or "").strip()
        else:
            flatten(child, key, record, columns)


def xml_to_csv(path, row_tag, out):
    root = parse(path).getroot()
    columns, records = [], []
    for row in root.findall(row_tag):
        record = {}
        flatten(row, "", record, columns)
        records.append(record)
    writer = csv.DictWriter(out, fieldnames=columns, restval="", extrasaction="ignore")
    writer.writeheader()
    writer.writerows(records)


# newline="" or csv writes \r\r\n on Windows.
# utf-8-sig writes the BOM Excel needs to read the file as UTF-8.
with open("out.csv", "w", newline="", encoding="utf-8-sig") as out:
    xml_to_csv(sys.argv[1], sys.argv[2], out)

import java.io.*;
import java.nio.charset.StandardCharsets;
import java.util.*;
import javax.xml.XMLConstants;
import javax.xml.parsers.*;
import org.w3c.dom.*;

public class XmlToCsv {

    static DocumentBuilder secureBuilder() throws Exception {
        DocumentBuilderFactory f = DocumentBuilderFactory.newInstance();
        // None of this is the default. Without it a document can read files
        // off the machine running the conversion.
        f.setFeature(XMLConstants.FEATURE_SECURE_PROCESSING, true);
        f.setFeature("http://apache.org/xml/features/disallow-doctype-decl", true);
        f.setFeature("http://xml.org/sax/features/external-general-entities", false);
        f.setFeature("http://xml.org/sax/features/external-parameter-entities", false);
        f.setXIncludeAware(false);
        f.setExpandEntityReferences(false);
        return f.newDocumentBuilder();
    }

    static void flatten(Element el, String prefix, Map<String, String> rec, List<String> cols) {
        NamedNodeMap attrs = el.getAttributes();
        for (int i = 0; i < attrs.getLength(); i++) {
            Node a = attrs.item(i);
            String key = prefix.isEmpty() ? "@" + a.getNodeName() : prefix + ".@" + a.getNodeName();
            if (!cols.contains(key)) cols.add(key);
            rec.put(key, a.getNodeValue());
        }

        List<Element> kids = new ArrayList<>();
        NodeList children = el.getChildNodes();
        for (int i = 0; i < children.getLength(); i++) {
            if (children.item(i) instanceof Element e) kids.add(e);
        }

        if (kids.isEmpty()) {
            String key = prefix.isEmpty() ? el.getNodeName() : prefix;
            if (!cols.contains(key)) cols.add(key);
            rec.put(key, el.getTextContent().trim());
            return;
        }
        for (Element k : kids) {
            String key = prefix.isEmpty() ? k.getNodeName() : prefix + "." + k.getNodeName();
            if (rec.containsKey(key)) {
                rec.put(key, rec.get(key) + "|" + k.getTextContent().trim());
            } else {
                flatten(k, key, rec, cols);
            }
        }
    }

    static String cell(String v) {
        if (v == null) return "";
        boolean needsQuotes = v.indexOf(',') >= 0 || v.indexOf('"') >= 0
                || v.indexOf('\r') >= 0 || v.indexOf('\n') >= 0;
        return needsQuotes ? '"' + v.replace("\"", "\"\"") + '"' : v;
    }

    public static void main(String[] args) throws Exception {
        Document doc = secureBuilder().parse(new File(args[0]));
        List<String> cols = new ArrayList<>();
        List<Map<String, String>> recs = new ArrayList<>();

        NodeList rows = doc.getDocumentElement().getElementsByTagName(args[1]);
        for (int i = 0; i < rows.getLength(); i++) {
            Map<String, String> rec = new LinkedHashMap<>();
            flatten((Element) rows.item(i), "", rec, cols);
            recs.add(rec);
        }

        try (PrintWriter out = new PrintWriter(new OutputStreamWriter(
                new FileOutputStream("out.csv"), StandardCharsets.UTF_8))) {
            out.print(String.join(",", cols.stream().map(XmlToCsv::cell).toList()) + "\r\n");
            for (Map<String, String> rec : recs) {
                out.print(String.join(",",
                        cols.stream().map(c -> cell(rec.get(c))).toList()) + "\r\n");
            }
        }
    }
}

using System.Text;
using System.Xml;
using System.Xml.Linq;

// A null XmlResolver means an external DTD or entity is never fetched.
var settings = new XmlReaderSettings
{
    DtdProcessing = DtdProcessing.Prohibit,
    XmlResolver = null,
    MaxCharactersFromEntities = 1024 * 1024,
};

using var reader = XmlReader.Create(args[0], settings);
var doc = XDocument.Load(reader);
var rowName = args[1];

var columns = new List<string>();
var records = new List<Dictionary<string, string>>();

void Flatten(XElement el, string prefix, Dictionary<string, string> rec)
{
    foreach (var a in el.Attributes())
    {
        if (a.IsNamespaceDeclaration) continue;   // xmlns is not data
        var key = prefix.Length == 0 ? "@" + a.Name.LocalName : prefix + ".@" + a.Name.LocalName;
        if (!columns.Contains(key)) columns.Add(key);
        rec[key] = a.Value;
    }

    var kids = el.Elements().ToList();
    if (kids.Count == 0)
    {
        var key = prefix.Length == 0 ? el.Name.LocalName : prefix;
        if (!columns.Contains(key)) columns.Add(key);
        rec[key] = el.Value.Trim();
        return;
    }

    foreach (var k in kids)
    {
        var key = prefix.Length == 0 ? k.Name.LocalName : prefix + "." + k.Name.LocalName;
        if (rec.ContainsKey(key)) rec[key] += "|" + k.Value.Trim();
        else Flatten(k, key, rec);
    }
}

foreach (var row in doc.Root!.Elements().Where(e => e.Name.LocalName == rowName))
{
    var rec = new Dictionary<string, string>();
    Flatten(row, "", rec);
    records.Add(rec);
}

static string Cell(string? v)
{
    if (v is null) return "";
    return v.IndexOfAny(new[] { ',', '"', '\r', '\n' }) >= 0
        ? "\"" + v.Replace("\"", "\"\"") + "\""
        : v;
}

// UTF8Encoding(true) writes a BOM, which is what makes Excel on Windows
// read the file as UTF-8 rather than the system code page.
using var writer = new StreamWriter("out.csv", false, new UTF8Encoding(true));
writer.WriteLine(string.Join(",", columns.Select(Cell)));
foreach (var rec in records)
{
    writer.WriteLine(string.Join(",", columns.Select(c => Cell(rec.GetValueOrDefault(c)))));
}

<?php
// LIBXML_NONET blocks network access during the parse. PHP 8 does not load
// external entities by default; passing the flag keeps this correct on PHP 7
// and documents the intent.
libxml_use_internal_errors(true);

$doc = new DOMDocument();
if (!$doc->load($argv[1], LIBXML_NONET)) {
    foreach (libxml_get_errors() as $e) {
        fprintf(STDERR, "line %d: %s", $e->line, $e->message);
    }
    exit(1);
}

$rowName = $argv[2];
$columns = [];
$records = [];

$flatten = function (DOMElement $el, string $prefix, array &$rec) use (&$flatten, &$columns) {
    foreach ($el->attributes as $a) {
        $key = $prefix === '' ? '@' . $a->name : $prefix . '.@' . $a->name;
        if (!in_array($key, $columns, true)) $columns[] = $key;
        $rec[$key] = $a->value;
    }

    $kids = [];
    foreach ($el->childNodes as $n) {
        if ($n instanceof DOMElement) $kids[] = $n;
    }

    if (!$kids) {
        $key = $prefix === '' ? $el->nodeName : $prefix;
        if (!in_array($key, $columns, true)) $columns[] = $key;
        $rec[$key] = trim($el->textContent);
        return;
    }

    foreach ($kids as $k) {
        $key = $prefix === '' ? $k->nodeName : $prefix . '.' . $k->nodeName;
        if (array_key_exists($key, $rec)) {
            $rec[$key] .= '|' . trim($k->textContent);
        } else {
            $flatten($k, $key, $rec);
        }
    }
};

foreach ($doc->documentElement->childNodes as $node) {
    if ($node instanceof DOMElement && $node->nodeName === $rowName) {
        $rec = [];
        $flatten($node, '', $rec);
        $records[] = $rec;
    }
}

$out = fopen('out.csv', 'w');
fwrite($out, "\xEF\xBB\xBF");        // BOM, for Excel
fputcsv($out, $columns);              // fputcsv applies RFC 4180 quoting
foreach ($records as $rec) {
    fputcsv($out, array_map(fn($c) => $rec[$c] ?? '', $columns));
}
fclose($out);

# xmlstarlet is the quickest route for a one-off extraction.
# --net=false stops it fetching a DTD or an external entity. Not the default.
xmlstarlet sel --net=false -t \
  -o 'id,title,price' -n \
  -m '/catalog/book' \
    -v '@id' -o ',' \
    -v 'title' -o ',' \
    -v 'price' -n \
  catalog.xml > out.csv

# It does no CSV quoting at all. A title containing a comma, a quote or a
# newline breaks the column alignment silently, so either be certain the
# data is clean or re-quote the output:
xmlstarlet sel --net=false -t -m '/catalog/book' \
  -v 'concat(@id,",",title)' -n catalog.xml | csvformat -   # csvkit

# For files too large to hold in memory, iterate instead of loading:
#   python -c 'from xml.etree.ElementTree import iterparse; ...'