Ok, got it now:) Sorry for the first e-mail, I got confused with this empty sequence vs. no argument thing. Goodnight. Lefteris On Thu, Feb 14, 2008 at 9:59 PM, Jan Rittinger wrote:

Lefteris,

sorry if my commit message was a bit misleading.

fn:name comes in two flavors fn:name() and fn:name(node()?). The first one is transformed into the second one during parse tree to core compilation. (This means we do not have to cope with the first function in the algebra.)

In my commit messages I wanted to refer to empty sequences (not missing parameters) like e.g. in

let $a := if (1=2) then <a/> else () return fn:name($a)

where the empty string (and not the empty sequence) needs to be returned.

Cheers, Jan

On 02/14/2008 05:31 PM, Torsten Grust wrote with possible deletions:

...

Lefteris,

I think Jan is talking an empty input _relation_ (in the loop lifting sense) not the absence of the argument in the actual function call.

Cheers, --Torsten

On Feb 14, 2008, at 15:37, Lefteris wrote:

...

Jan, is this the correct way to cope with empty input?

The xquery function specs say that if input is ommited then it defaults to the context item (.)

That also is expresed in xquery_fo.c with the .arity = 0 version of fn:name etc. If that is true, shouldn't be dealed at the core level?

lefteris

On Wed, Feb 13, 2008 at 9:54 PM, Jan Rittinger wrote:

...

Update of /cvsroot/monetdb/pathfinder/compiler/algebra In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv18492/algebra

Modified Files: builtins.c Log Message: -- Fix the translation of fn:name, fn:local-name, and fn:namespace- uri. (All three have to cope with empty input.)

Index: builtins.c =================================================================== RCS file: /cvsroot/monetdb/pathfinder/compiler/algebra/builtins.c,v retrieving revision 1.79 retrieving revision 1.80 diff -u -d -r1.79 -r1.80 --- builtins.c 6 Feb 2008 16:36:28 -0000 1.79 +++ builtins.c 13 Feb 2008 20:54:17 -0000 1.80 @@ -2350,6 +2350,10 @@ * is not even the right function to use there, because the constructors * should do their job based on _statically-known namespace declarations_, * not in-scope declarations. + * + * Furthermore a cast from string to xs:QName is only allowed at compile + * time on constant strings. At runtime a cast to xs:QName is only allowed + * for QNames inputs. */ struct PFla_pair_t PFbui_fn_resolve_qname (const PFla_op_t *loop, bool ordering, @@ -2416,22 +2420,34 @@ struct PFla_pair_t PFbui_fn_name (const PFla_op_t *loop, bool ordering, struct PFla_pair_t *args) { - (void) loop; (void) ordering; + (void) ordering;

- return (struct PFla_pair_t) { - .rel = project ( - fun_1to1 ( - project (args[0].rel, - proj (att_iter, att_iter), - proj (att_pos, att_pos), - proj (att_item, att_item)), - alg_fun_fn_name, - att_res, - attlist(att_item)), - proj (att_iter, att_iter), - proj (att_pos, att_pos), - proj (att_item, att_res)), - .frag = PFla_empty_set() }; + PFla_op_t *strings = project ( + fun_1to1 ( + project (args[0].rel, + proj (att_iter, att_iter), + proj (att_pos, att_pos), + proj (att_item, att_item)), + alg_fun_fn_name, + att_res, + attlist(att_item)), + proj (att_iter, att_iter), + proj (att_pos, att_pos), + proj (att_item, att_res)); + + PFla_op_t *res = disjunion ( + strings, + attach ( + attach ( + difference ( + loop, + project ( + strings, + proj (att_iter, att_iter))), + att_pos, lit_nat (1)), + att_item, lit_str (""))); + + return (struct PFla_pair_t) { .rel = res, .frag = PFla_empty_set() }; }

/* ------------------- */ @@ -2442,22 +2458,34 @@ PFbui_fn_local_name (const PFla_op_t *loop, bool ordering, struct PFla_pair_t *args) { - (void) loop; (void) ordering; + (void) ordering;

- return (struct PFla_pair_t) { - .rel = project ( - fun_1to1 ( - project (args[0].rel, - proj (att_iter, att_iter), - proj (att_pos, att_pos), - proj (att_item, att_item)), - alg_fun_fn_local_name, - att_res, - attlist(att_item)), - proj (att_iter, att_iter), - proj (att_pos, att_pos), - proj (att_item, att_res)), - .frag = args[0].frag }; + PFla_op_t *strings = project ( + fun_1to1 ( + project (args[0].rel, + proj (att_iter, att_iter), + proj (att_pos, att_pos), + proj (att_item, att_item)), + alg_fun_fn_local_name, + att_res, + attlist(att_item)), + proj (att_iter, att_iter), + proj (att_pos, att_pos), + proj (att_item, att_res)); + + PFla_op_t *res = disjunion ( + strings, + attach ( + attach ( + difference ( + loop, + project ( + strings, + proj (att_iter, att_iter))), + att_pos, lit_nat (1)), + att_item, lit_str (""))); + + return (struct PFla_pair_t) { .rel = res, .frag = PFla_empty_set() }; }

/* --------------------- */ @@ -2468,22 +2496,34 @@ PFbui_fn_namespace_uri (const PFla_op_t *loop, bool ordering, struct PFla_pair_t *args) { - (void) loop; (void) ordering; + (void) ordering;

- return (struct PFla_pair_t) { - .rel = project ( - fun_1to1 ( - project (args[0].rel, - proj (att_iter, att_iter), - proj (att_pos, att_pos), - proj (att_item, att_item)), - alg_fun_fn_namespace_uri, - att_res, - attlist(att_item)), - proj (att_iter, att_iter), - proj (att_pos, att_pos), - proj (att_item, att_res)), - .frag = args[0].frag }; + PFla_op_t *strings = project ( + fun_1to1 ( + project (args[0].rel, + proj (att_iter, att_iter), + proj (att_pos, att_pos), + proj (att_item, att_item)), + alg_fun_fn_namespace_uri, + att_res, + attlist(att_item)), + proj (att_iter, att_iter), + proj (att_pos, att_pos), + proj (att_item, att_res)); + + PFla_op_t *res = disjunion ( + strings, + attach ( + attach ( + difference ( + loop, + project ( + strings, + proj (att_iter, att_iter))), + att_pos, lit_nat (1)), + att_item, lit_str (""))); + + return (struct PFla_pair_t) { .rel = res, .frag = PFla_empty_set() }; }

/* --------------- */

------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Monetdb-pf-checkins mailing list Monetdb-pf-checkins@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-pf-checkins

------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Monetdb-developers mailing list Monetdb-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-developers

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

------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Monetdb-developers mailing list Monetdb-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-developers