如何使用lxml XPath匹配指定命名空间（如cims与cims51）而非全部命名空间

Great question! Unfortunately, native XPath doesn’t support wildcard syntax like cims*: for matching namespace prefixes directly. But there are two solid workarounds to handle your scenario, depending on whether you want to match based on the namespace URI (more reliable) or the prefix name (convenient if you control the prefix mappings):

Option 1: Match by Namespace URI (Recommended) #

Since all your cims/cims51 prefixes map to URIs starting with urn:iec62325.351:tc57wg16:451-2:scheduledocument:, you can use starts-with() on the element's namespace URI to target exactly those mRID nodes, regardless of their prefix. This is more robust because prefixes are just aliases—XML documents could use different prefixes for the same standard URI, but this approach still works.

Use this XPath expression:

mrids = xml.xpath(
    r"//*[local-name()='mRID' and starts-with(namespace-uri(), 'urn:iec62325.351:tc57wg16:451-2:scheduledocument:')]",
    namespaces=NAMESPACES
)

Option 2: Dynamically Generate XPath from Prefixes #

If you want to strictly match prefixes starting with cims (and you’ve already mapped all relevant prefixes in your NAMESPACES dictionary), you can generate the XPath expression programmatically in Python. This avoids manually listing every possible cims variant.

Here’s how to do it:

# Filter all prefixes in NAMESPACES that start with "cims"
cims_prefixes = [prefix for prefix in NAMESPACES if prefix.startswith("cims")]

# Build the XPath expression by joining all prefix-specific mRID selectors
xpath_expr = " | ".join([f"//{prefix}:mRID" for prefix in cims_prefixes])

# Execute the XPath query
mrids = xml.xpath(xpath_expr, namespaces=NAMESPACES)

Key Notes: #

• The cims*: syntax isn’t valid in XPath because prefixes are treated as literal identifiers, not patterns. The * wildcard only works for element/attribute names, not prefixes.
• Option 1 is preferred for production code because it relies on the standard namespace URI (the actual identifier of the namespace) rather than arbitrary prefix aliases.

内容的提问来源于stack exchange，提问作者MortenB