Nicolai Tufar wrote:
> From: "Tom Lane" <tgl@sss.pgh.pa.us>
>>Hm. Offhand it would seem that somebody's output function is returning
>>a non-palloc'd string. What are the datatypes in that table, exactly?
>
> apb=> \d maras2.mrk_yyn
> Table "maras2.mrk_yyn"
> Column | Type | Modifiers
> ---------+-----------------------+-----------
> bsvr_no | character varying(10) |
> yyn_tur | character varying(1) |
> yyn_no | numeric |
> yyn_syf | numeric |
> yyn_trh | date |
FWIW, I just tried to repeat this on cvs tip and could not. Seems to work just
fine:
regression=# create table foo(bsvr_no character varying(10),yyn_tur character
varying(1),yyn_no numeric,yyn_syf numeric,yyn_trh date);
CREATE TABLE
regression=# insert into foo values('a','a','1.0','1.0',now()::date);
INSERT 556825 1
regression=# copy foo to stdout;
a a 1.0 1.0 2002-11-30
regression=# copy foo(bsvr_no) to stdout;
a
regression=# copy foo(bsvr_no,yyn_tur,yyn_no,yyn_syf,yyn_trh) to stdout;
a a 1.0 1.0 2002-11-30
regression=# insert into foo values('a','a','1.0','1.0',null);
INSERT 556826 1
regression=# copy foo(bsvr_no,yyn_tur,yyn_no,yyn_syf,yyn_trh) to stdout;
a a 1.0 1.0 2002-11-30
a a 1.0 1.0 \N
Joe
