Python如何提取含XML处理指令的非标准XML中的完整坐标数据？

我用Python从非标准XML文件提取坐标数据，部分<coordinates>标签里包含XML处理指令，示例XML如下：

<?xml version="1.0" encoding="UTF-8"?>
<kml>
    <Document>
        <Folder>
            <Placemark id="CAR1">
            <coordinates>
                28.236931382585154893,36.380279408060658852,200<?AtTime 0.000000?>
                28.236862947306430982,36.380324400967296583,200<?AtTime 2.333333?>
                28.236862521767100986,36.380323843439811071,200<?AtTime 2.500000?>
            </coordinates>
                </Placemark>
        <Placemark id="CAR2">
            <coordinates>
                28.236931382585154893,36.380279408060658852,200
                28.236862947306430982,36.380324400967296583,200
                28.236862521767100986,36.380323843439811071,200
            </coordinates>
        </Placemark>
       </Folder>
    </Document>
</kml>

我的测试代码：

from xml.dom import minidom
myFile = "test.xml"
mydoc = minidom.parse(myFile)
items = mydoc.getElementsByTagName('Placemark')
for elem in items:
    print("coordinates ",elem.attributes['id'].value)
    coorTag = elem.getElementsByTagName('coordinates')
    if len(coorTag) > 0:
        for elem in coorTag:
            theData = elem.firstChild.nodeValue
            print(theData)

当前输出：

coordinates  CAR1

                28.236931382585154893,36.380279408060658852,200
coordinates  CAR2

                28.236931382585154893,36.380279408060658852,200
                28.236862947306430982,36.380324400967296583,200
                28.236862521767100986,36.380323843439811071,200

期望输出：

coordinates  CAR1

                28.236931382585154893,36.380279408060658852,200<?AtTime 0.000000?>
                28.236862947306430982,36.380324400967296583,200<?AtTime 2.333333?>
                28.236862521767100986,36.380323843439811071,200<?AtTime 2.500000?>
coordinates  CAR2

                28.236931382585154893,36.380279408060658852,200
                28.236862947306430982,36.380324400967296583,200
                28.236862521767100986,36.380323843439811071,200

请问如何获取包含XML处理指令的<coordinates>标签内的完整坐标数据？这是否是minidom的问题，有没有更合适的Python替代库？

1. 修复minidom的处理逻辑 #

minidom会把XML处理指令（PI）识别为独立节点，而非文本节点的一部分，所以只取firstChild.nodeValue只能拿到第一个文本节点的内容，后续的处理指令和文本节点都会被忽略。

要获取<coordinates>内的完整内容，需要遍历该标签下的所有子节点，根据节点类型拼接内容：

• 文本节点（nodeType == Node.TEXT_NODE）：直接取nodeValue
• 处理指令节点（nodeType == Node.PROCESSING_INSTRUCTION_NODE）：拼接成<?target data?>格式的字符串

修改后的代码：

from xml.dom import minidom
from xml.dom.minidom import Node

myFile = "test.xml"
mydoc = minidom.parse(myFile)
items = mydoc.getElementsByTagName('Placemark')

for elem in items:
    print("coordinates ", elem.attributes['id'].value)
    coorTag = elem.getElementsByTagName('coordinates')
    if len(coorTag) > 0:
        for coor_elem in coorTag:
            full_content = []
            for child in coor_elem.childNodes:
                if child.nodeType == Node.TEXT_NODE:
                    full_content.append(child.nodeValue)
                elif child.nodeType == Node.PROCESSING_INSTRUCTION_NODE:
                    pi_str = f"<?{child.target} {child.data}?>"
                    full_content.append(pi_str)
            print(''.join(full_content))

这段代码会遍历<coordinates>下的所有子节点，将文本和处理指令完整拼接，得到预期结果。

2. 更合适的替代库 #

如果经常处理这类非标准XML，推荐使用lxml库，它对XML节点的处理更灵活，能直接获取节点原始内容：

安装lxml #

pip install lxml

使用lxml的代码示例 #

from lxml import etree

myFile = "test.xml"
tree = etree.parse(myFile)
placemarks = tree.xpath('//Placemark')

for pm in placemarks:
    pm_id = pm.get('id')
    print(f"coordinates  {pm_id}")
    coor_elem = pm.find('coordinates')
    if coor_elem is not None:
        full_content = etree.tostring(coor_elem, encoding='unicode', method='text')
        print(full_content)

lxml的etree.tostring方法通过method='text'参数，可直接提取节点内的所有内容（包括处理指令），无需手动遍历节点，代码更简洁。

另外，BeautifulSoup也可将XML当作HTML解析处理，但lxml在XML领域的专业性更强。

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