On Tue, Dec 8, 2009 at 5:43 AM, Teemu Juntunen <teemu.juntunen@xxxxxxxxxx> wrote: > Hello, > > is there any possibility to get null-values as text 'null' from > dynamic select expression like > > SELECT ROW(t.*) FROM table t > > Normally you get > > (t1,t2,t3,,t5,,,) > > How can you get > > (t1,t2,t3,null,t5,null,null,null) > > I would like to use the result with dynamic update expression and update > doesn't like of empty values, so > > UPDATE SET (t1,t2,t3,t4,t5,t6,t7,t8) = (t1,t2,t3,,t5,,,) It can't be done. The problem is that the update statement has a funky syntax which is not composite type friendly. What I would personally like to be table to do is: update foo set foo = (x,y,z)::foo where... then you could do: update foo set foo = '(a,b,)'::foo where... Couple of posssible workarounds: *) If you are willing to put the fields you are updating as a block into a composite type, then you could update them as a block: create table bar as (...) create table foo (b bar, ...) update foo set bar = ()::bar where...; This isn't a general solution obviously. *) convert your update to insert+delete: delete from foo where... insert into foo select '(a,b,)'::foo; watch that race condition! merlin -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
