问题
I need to compare XML values with PostgreSQL 9.X. Example:
SELECT 1 FROM t WHERE xml1=xml2; -- error for XML datatype
but can do only with casting to text datatype,
SELECT 1 FROM t WHERE xml1::text=xml2::text; -- OK, but is a text-comparison
that is not the best way to do this. I need a better comparison algorithm... Looking for it, I find that, perhaps, "Canonical XML" can be compared.
How to do "Canonical XML comparison" with PostgreSQL?
NOTE: PostgreSQL team assumes false premisse about XML comparison?
Many documents about XML comparison sugest that Canonical XML is the reference format for that operation. Ex. at Wikipedia they say,
According to the W3C, if two XML documents have the same canonical form, then the two documents are logically equivalent within the given application context.
But, PostgreSQL docs/9.2 says,
The xml data type is unusual in that it does not provide any comparison operators. This is because there is no well-defined and universally useful comparison algorithm for XML data.
Well, for me it is a false premise, and can be hiden a real lack of an important PostgreSQL issue.
回答1:
I contributed to the XML type implementation in PostgreSQL and probably wrote most of the documentation you are citing.
There are a number of reasons why this is currently the way it is:
- The SQL standard specifies no comparison operator for type
xml. - At the time the implementation was started, Canonical XML wasn't widely used and understood (at least by the people involved, arguably).
- There are certain limitations where XML canonicalization doesn't work. Although these might be rarely seen in practice, this would yield the situation where some values of a data type can't be compared, which would lead to problems with indexing for example. (The NaN value of floating-point types is assigned an ordering position for similar reasons.)
- It's still debatable whether comparison by canonicalization is appropriate for all uses and what users always want.
An implementation of an XML canonicalization function for optional use would certainly be welcome. I would actually like to see a separate xmlcanonical type, but that would be quite a bit more work.
来源：https://stackoverflow.com/questions/15393026/how-to-do-xml-canonical-comparison-with-postgresql