Re: 9th's bug

I find this very interesting : EXECUTE PROCEDURE p( ( SELECT f FROM t2 ) ); And the argument to p is a nvarchar(10)... What is "rotor" trying to do? What about if the select on t2 returns more than one row? J. Jonathan Leffler wrote: > rotor wrote: > >>Hi community! >> >>IDS 9.40.TC6 >> >>CREATE TABLE t1 ( f NVARCHAR( 10 ) ); >>CREATE TABLE t2 ( f NVARCHAR( 10 ) NOT NULL ); >>CREATE PROCEDURE p( p CHAR( 10 ) ) END PROCEDURE;>> >>EXECUTE PROCEDURE p( ( SELECT f FROM t2 ) );>>-- Oops! SQL Error (-674) : Routine (p) can not be resolved. >> >>But if the folowed SQL will be executed first: >>EXECUTE PROCEDURE p( ( SELECT f FROM t1 ) );>>then the problematic statement also works properly. Funny? >> >>Any comments awaiting... > > > > It sounds suspiciously like: > B172186 EXECUTE PROCEDURE STATEMENT RETURNS 674 ERROR WHEN RUN FIRST TIME. > > sending to informix-list

Jean Sagi wrote: > I find this very interesting : > > EXECUTE PROCEDURE p( ( SELECT f FROM t2 ) );> > And the argument to p is a nvarchar(10)... > > What is "rotor" trying to do? > > What about if the select on t2 returns more than one row? > > J. Should be the same error as when you do: 1 + (select f from t2) At the end of the day (select f from t2) is just a (scalar) expression. I presume if t2 has no rows the result of the scalar subquery is NULL::NVARCHAR(10)? Cheers Serge Serge Rielau DB2 SQL Compiler Development IBM Toronto Lab