This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Invoking fn:put on the same URI twice or more, with nodes that have the same node identity or not, is currently not discussed in the 2007/August draft. This topic has been raised twice on the w3c-xml-query-wg@w3.org list, by John Snelson "Multiple invocations of fn:put() with the same URI"(2007-06-06) and by me, "fn:put: multiple invocations on the same URI"(2007-11-27). Member-only links: http://lists.w3.org/Archives/Member/w3c-xml-query-wg/2007Jun/0006.html http://lists.w3.org/Archives/Member/w3c-xml-query-wg/2007Nov/0079.html One way to implement it would be to let fn:put be an update primitive(it's already required to be invoked at the end of the query, for instance). Report #4167 is related.
On 15 Jan 2008 the working group considered this bug report. Since the semantics of fn:put() are implementation-defined, the group decided that it would be inappropriate to specify its error behavior. Error conditions, like the rest of fn:put() semantics, can be defined by the implementation. Frans, if you are satisfied with this resolution, please mark this bug report as "Closed". --Don Chamberlin (for the Query Working Group)
I would appreciate some insight into why this particularly area wasn't made to not be implementation defined. The gain this behavior has is that implementations can define something meaningful of having multiple invocations on a URI. Since the order of those invocations is undefined/arbitrary, I wonder what possible usage scenarios this could have. The cost is, of course, reduced interoperability.
Since we do not define the external semantics of fn:put() when invoked once in a snapshot, we can't really clarify the external semantics of fn:put() when invoked more than once. Jonathan Robie (for the XQuery Working Group)