Hi Peter, your checkin broke the implementation of fn:id() (see the testweb for the tests in benchmarks/XPath and benchmarks/XPathMark). Jan On 10/26/2006 02:12 PM, Peter Boncz wrote with possible deletions:

Update of /cvsroot/monetdb/pathfinder/runtime In directory sc8-pr-cvs7.sourceforge.net:/tmp/cvs-serv8300/runtime

Modified Files: pathfinder.mx pf_support.mx Log Message: fast NID node retriveal by NID - moved support for id/idref node retrieval into pathfinder.mx having this as an API function will allow us to provide this functionality using indices later (i.e. be able to lock and isolte a hash index). it also helps a port of this functionality to algebra.

for the moment, there are no shared hash tables, and ID/IDREF queries on updatable XML documents will be slowish (sequential scans needed) the new NID access is a good alternative to look up nodes.

- implemented in ID node retrieval support for string sthat are numbers, these are interpreted as NIDs, not ID attributes.

- implemented a new fn:nid(element()*) as xs:string* builtin function that returns the NIDs of element nodes.

- fixed some constant support things (mposjoin)

[...]

-- Jan Rittinger Database Systems Technische Universität München (Germany) http://www-db.in.tum.de/~rittinge/