Re: Errors when restoring backup created by pg_dumpall

Adrian Klaver <adrian(dot)klaver(at)aklaver(dot)com> writes:
> On 12/1/24 13:14, Tom Lane wrote:
>> It would be useful to know what is the command at line 4102
>> of all.sql.

> It is here:

> https://gist.github.com/poperigby/fcb59eb6c22c6051800e06a0ec482b49

> CREATE TABLE public.geodata_places (
> id integer NOT NULL,
> name character varying(200) NOT NULL,
> longitude double precision NOT NULL,
> latitude double precision NOT NULL,
> "countryCode" character(2) NOT NULL,
> "admin1Code" character varying(20),
> "admin2Code" character varying(80),
> "modificationDate" date NOT NULL,
> "earthCoord" public.earth GENERATED ALWAYS AS
> (public.ll_to_earth(latitude, longitude)) STORED,
> "admin1Name" character varying,
> "admin2Name" character varying,
> "alternateNames" character varying
> );

Ah! Then the failure occurs because we do a planning pass on the
GENERATED expression (I don't remember exactly why that's needed
during CREATE TABLE). So maybe messing with the dump script's
search_path setting *would* be enough to get you past that.

Having said that, the CREATE should have been seeing the new-style
definition of ll_to_earth() if the 1.2 version of earthdistance
was correctly installed.

regards, tom lane