Re: Is this ok with Informix?

Madison Pruet wrote: > I'm sure that Jonathan is not implying that IDS does not support cascade > deletes as in ... Actually I was - but I was misthinking and hence mistaken. :-( It is the cascade update that isn't supported. And ON DELETE RESTRICT is also not supported by explicit syntax, as I said originally. And while I was checking my mistake, I looked at the manual (c1sqna.pdf for 9.40 Guide to SQL: Syntax) and noted that on p2-214, the railroad diagram clearly implies that all multi-column constraints must follow all the column definitions, whereas I'm 99% sure that you can interleave them as long as any columns referenced by the multi-column constraint have already been defined. CREATE TABLE pqr ( s integer not null check (s > 0), t integer not null check(t < 0), primary key (s, t) constraint pk_pqr, u integer not null ); > Create table tab1 ( > col1 integer primary key, > ... > )> > Create table tab2 (> col1 char(10) primary key > col2 integer references tab1 (col1) ON DELETE CASCADE > ); > > which IDS does support... > > "Jonathan Leffler" <jleffler@earthlink.net> wrote: >>Carlos San Miguel wrote: >>>Is this syntax legal in Informix? >>> >>>ALTER TABLE person ADD CONSTRAINT FOREIGN KEY(cnr) REFERENCES comp(cnr) >>>ON DELETE RESTRICT;>> >>No - as you must realize since you wouldn't be asking if you hadn't >>gotten a syntax error from running the statement. >> >>Download the manuals from: >> >>http://www.ibm.com/software/data/informix/library >> >>Informix does not support the explicit 'ON DELETE RESTRICT' notation, >>but does support the concept -- indeed, it does not support any >>alternatives, unless you program them through triggered code. -- Jonathan Leffler (jleffler@earthlink.net, jleffler@us.ibm.com) #include <disclaimer.h> Guardian of DBD::Informix v2003.04 -- http://dbi.perl.org/