On 01/28/25 03:14, Jim Jones wrote:
> I'd say the difference is due to how the two systems handle the XML data
> type and unquoted identifiers in general, rather than a difference in
> the behaviour of the function itself.
I'd go further and say it's entirely down to how the two systems
handle unquoted identifiers. In neither case was there ever any
XML value created with XML names in one case and then changed to
the other. The SQL names were already in their (DB2- or PostgreSQL-
specific) folded form by the first moment any XML library code ever
saw them. The XML code handled them faithfully ever after, whether
in serialized or in node-tree form.
Presumably both DB2 and PostgreSQL users soon know in their sleep
what their respective systems do to unquoted identifiers, and know
that quoting is the way to control that when it matters.
Regards,
-Chap
