在 Python 中将 XML 解析为 CSV

我按照 python 文档处理 Unicode。我首先检查了编码。XML声明似乎不正确：errors="surrogateescape"charset_normalizer

import xml.etree.ElementTree as ET
from charset_normalizer import detect
import csv

filename = 'Gemeentelijk Gegevensmodel XMI2.1.2.xml'

with open(filename, 'rb') as f:
    xml = f.read()

result = detect(xml)
print(result)
for k, v in result.items():
    if k == "confidence":
        en_conf = v
        print(k, v)
    if k == "encoding":
        en_cod = v
        print(k, v)

if en_conf == 1.0:
    with open(filename, 'r', encoding=en_cod, errors="surrogateescape") as f:
        data = f.read()
        # remove wrong declaration
        data1 = data.replace("<?xml  version='1.0' encoding='windows-1252' ?>\n","")
   
    with open('NewXML_'+filename, 'w', encoding="utf-8", errors="surrogateescape") as f:
        # add new declaration
        f.write("<?xml  version='1.0' encoding='UTF-8' ?>\n")
        f.write(data1)
        
     
tree = ET.parse('NewXML_'+filename)
root = tree.getroot()

ns = dict([node for _, node in ET.iterparse('NewXML_'+filename, events=['start-ns'])])
#print(ns)

table_data = []
packagedElements = root.findall('.//packagedElement[@xmi:type="uml:Class"]', ns)
print(len(packagedElements))
for elem in packagedElements:
    if elem.find(".//ownedAttribute") is not None and elem.find(".//ownedAttribute").get("name") is not None:
        name = elem.find(".//ownedAttribute").get("name")     
    if elem.find(".//type") is not None and elem.find(".//type").get("{http://schema.omg.org/spec/XMI/2.1}idref") is not None:
        ty = elem.find(".//type").get("{http://schema.omg.org/spec/XMI/2.1}idref")

    row = name, ty

    table_data.append(row)

print(table_data)
print(len(table_data))

with open("table.csv", "w", encoding='utf-8', newline="\n") as csv_file:
    c_writer = csv.writer(csv_file, delimiter=';', quoting=csv.QUOTE_MINIMAL)
    for row in table_data:
        c_writer.writerow(row)

输出：

{'encoding': 'cp775', 'language': 'English', 'confidence': 1.0}
encoding cp775
confidence 1.0
742

[('contactsoort', 'EAJava_AN80'), ('aantal kamers', 'EAJava_int'), ('einddatum', 'EAJava_Date'), ('winkelformule', 'EAJava_AN80'), ('aantalFulletimeVrouwen', 'EAJava_int'), ('aantalkassa', 'EAJava_int'), ('aanvangAanwezigheid', 'EAJava_DateTime'), ('nummer', 'EAJava_int'), ('naam', 'EAJava_AN250'), ('aanstellingsdatum', 'EAJava_Date'), ('naam', 'EAJava_AN250'), ('naam', 'EAJava_AN250'), ('naam', 'EAJava_AN250'), ('naam', 'EAJava_AN250'), ('aanstellingsdatum', 'EAJava_Date'), ('besloten', 'EAJava_Boolean'), ... shorted ]

742

我试图分析XML并查看哪些行（单词）的字节超过127（\x7F）（“高位”）：

import csv

with (
    open("input.xml", "rb") as f_in,
    open("output.csv", "w", newline="") as f_out,
):
    writer = csv.writer(f_out)
    writer.writerow(["Line #", "Words w/high bits"])

    i, line = int, bytes
    for i, line in enumerate(f_in, start=1):
        line = line.replace(b"\x09", b"").replace(b"\x0a", b"")  # drop tabs and newlines
        words: list[bytes] = []
        for word in line.split(b" "):
            for x in word:
                if x > 127:
                    words.append(word)
                    break
        if words != []:
            writer.writerow([i, words])

我在这两行中发现了“Plaatsoriëntatie”一词的不一致编码：

Line #, Words w/high bits
...
  7638, "[b'name=""Plaatsori\xc3\xabntatie"">']"
  7641, "[b'xmi:idref=""EAJava_enum_Plaatsori\xc3_ntatie""/>']"
...

• 第一行正确地将“ë”编码为 UTF-8 （）。\xc3\xab
• 第二行看起来曾经被正确地 UTF-8 编码，然后（在 ISO-8859-1 和 Windows-1252 中为“«”）看起来已被替换为“_”。\xab

我有一种预感，源具有格式错误的编码（例如，）;如果你能深入了解它，我很想了解它的真相。祝你好运。\xc3_