UDDF XML转CSV的Python脚本优化：正则或ElementTree方案咨询

我写了一段将UDDF XML文件转换为CSV格式的Python脚本，目前用拆分浮点数后重新拼接的方式处理数值，方法比较粗糙。想请教：仅用正则表达式，或者借助ElementTree库，有没有更优的实现方案？

我的Python脚本 #

import re 

uddf_file = open("c:/temp/python/uddf_to_csv/input.xml", "r")
csv_file = open("c:/temp/python/uddf_to_csv/output.csv", "x")

while True:

    uddf_line = uddf_file.readline()
    uddf_line.rstrip('\r\n')

    depth_val = re.search("depth", uddf_line)
    if depth_val: depth_val_num = re.findall(r'\d+', uddf_line)
    if depth_val: csv_file.write(depth_val_num[0] + "." + depth_val_num[1] + ",")

    time_val = re.search("divetime", uddf_line)
    if time_val: time_val_num = re.findall(r'\d+', uddf_line)
    if time_val: csv_file.write(time_val_num[0] + "." + time_val_num[1] + ",")

    temp_val = re.search("temperature", uddf_line)
    if temp_val: temp_val_num = re.findall(r'\d+', uddf_line)
    if temp_val: csv_file.write(temp_val_num[0] + "." + temp_val_num[1] + "\n")   

    if not uddf_line: break;

uddf_file.close()
csv_file.close()

示例UDDF文件 #

<?xml version="1.0" encoding="utf-8"?>
<uddf xmlns="http://www.streit.cc/uddf/3.2/" version="3.2.0">
  <generator>
    <name>awu2uddf</name>
    <manufacturer id="Foghead">
      <name>Doug Junkins</name>
    </manufacturer>
    <version>1.1(3)</version>
  </generator>
  <diver>
    <owner id="owner">
      <personal>
        <firstname/>
        <lastname/>
      </personal>
      <equipment>
        <divecomputer id="81333b11">
          <name>Apple Watch Ultra</name>
           <model>Apple Watch Ultra</model>
        </divecomputer>
      </equipment>
    </owner>
    <buddy/>
  </diver>
  <divesite/>
  <gasdefinitions/>
  <profiledata>
    <repetitiongroup id="2023-04-12T17:02:08Z">
      <dive id="2023-04-12T17:02:08Z">
        <informationbeforedive>
          <datetime>2023-04-12T10:02:08-0700</datetime>
        </informationbeforedive>
        <samples>
          <waypoint>
            <depth>2.086</depth>
            <divetime>2.520</divetime>
            <temperature>293.3</temperature>
          </waypoint>
          <waypoint>
            <depth>3.322</depth>
            <divetime>5.150</divetime>
            <temperature>293.3</temperature>
          </waypoint>
          <waypoint>
            <depth>1.101</depth>
            <divetime>2263.999</divetime>
            <temperature>293.2</temperature>
          </waypoint>
        </samples>
      </dive>
    </repetitiongroup>
  </profiledata>
</uddf>

方案一：改进正则表达式实现 #

原脚本拆分浮点数再拼接的方式容错性差，遇到整数或跨行数值会失效。直接匹配完整数值的正则方案更可靠：

import re

# 用with语句自动管理文件，避免手动关闭遗漏
with open("c:/temp/python/uddf_to_csv/input.xml", "r") as uddf_file, open("c:/temp/python/uddf_to_csv/output.csv", "x") as csv_file:
    current_row = []
    # 匹配标签和对应数值，支持整数、浮点数格式
    pattern = re.compile(r'<(depth|divetime|temperature)>(\d+\.?\d*|\.\d+)</\1>')
    
    for line in uddf_file:
        matches = pattern.findall(line)
        for tag, value in matches:
            current_row.append(value)
            # 收集到三个字段后写入CSV，确保属于同一个waypoint
            if len(current_row) == 3:
                csv_file.write(','.join(current_row) + '\n')
                current_row = []

核心改进 #

• 用正则直接提取完整数值，无需拆分拼接，兼容整数和浮点数
• 新增临时列表current_row，避免字段错位
• 使用with语句自动处理文件生命周期，代码更简洁安全

方案二：使用ElementTree库（更推荐） #

XML是结构化数据，用专门解析库比正则更稳定，尤其能处理UDDF的命名空间和嵌套结构：

import xml.etree.ElementTree as ET
import csv

# 定义UDDF的命名空间，XML中必须指定才能正确查找节点
NS = {'uddf': 'http://www.streit.cc/uddf/3.2/'}

# 解析XML文件
tree = ET.parse("c:/temp/python/uddf_to_csv/input.xml")
root = tree.getroot()

# 写入CSV文件
with open("c:/temp/python/uddf_to_csv/output.csv", "x", newline='') as csv_file:
    writer = csv.writer(csv_file)
    # 遍历所有waypoint节点
    for waypoint in root.findall('.//uddf:waypoint', NS):
        # 提取每个节点的字段值
        depth = waypoint.find('uddf:depth', NS).text
        divetime = waypoint.find('uddf:divetime', NS).text
        temperature = waypoint.find('uddf:temperature', NS).text
        # 写入一行数据
        writer.writerow([depth, divetime, temperature])

显著优势 #

• 完全贴合XML结构解析，不会出现正则的误匹配问题
• 自动处理UDDF的命名空间，原正则脚本未处理此情况，若标签带命名空间会失效
• 使用csv.writer模块规范写入CSV，自动处理潜在的特殊字符问题
• 代码逻辑清晰，直接按节点层级提取数据，维护成本低

内容的提问来源于stack exchange，提问作者IronDevil74