Re: [HACKERS] Re: Changing the default value of an inherited column

[Tom Lane]

Philip Warner <[EMAIL PROTECTED]> writes:
> Not a squawk as such, but does this have implications for pg_dump?

Good point.  With recently-committed changes, try:

regression=# create table p1 (f1 int default 42 not null, f2 int);
CREATE
regression=# create table c1 (f1 int, f2 int default 7) inherits (p1);
NOTICE:  CREATE TABLE: merging attribute "f1" with inherited definition
NOTICE:  CREATE TABLE: merging attribute "f2" with inherited definition
CREATE
regression=# create table c2 (f1 int default 43, f2 int not null) inherits (p1);
NOTICE:  CREATE TABLE: merging attribute "f1" with inherited definition
NOTICE:  CREATE TABLE: merging attribute "f2" with inherited definition
CREATE

pg_dump dumps both c1 and c2 like this:

CREATE TABLE "c2" (

)
inherits ("p1");

which is OK as far as the field set goes, but it loses the additional
DEFAULT and NOT NULL information for the child table.  Any thoughts on
the best way to fix this?

                        regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]