On Tue, Oct 6, 2026 at 3:52 PM Fujii Masao <[email protected]> wrote: > > So, if two source tables both set fillfactor, wouldn't > > CREATE TABLE ... (LIKE t1 INCLUDING ALL, LIKE t2 INCLUDING ALL) > > fail with this patch even though it succeeds today? I'm concerned that this > could be a backward compatibility issue. >
Hi. CREATE TABLE LIKE is not guaranteed to be error-free. For example, the following will fail: create table t(a int); create table t1(a int); CREATE TABLE t2 (LIKE t INCLUDING ALL, LIKE t1 INCLUDING ALL); I think this failure is not different from the fillfactor one: certain conditions have to be met, otherwise CREATE TABLE ... LIKE fails.
