Inserts using plpgsql - Further

From: "Scott Holmes" <sholmes@pacificnet.net>

Well, it does not seem to be a problem with multiple fields but with

fields

that are integers.

INSERT INTO journal (j_date,casenmbr,j_descr,j_rate,j_status) VALUES
(jdate,cnmbr,erec.ev_title,erec.ev_rate,jstatus);

Fails, but

INSERT INTO journal (j_date,j_descr,j_rate,j_status) VALUES
(jdate,erec.ev_title,erec.ev_rate,jstatus);

Succeeds. The difference is casenmbr, an integer. j_rate is a decimal

value,

j_date is a date and the others are character fields.

Puzzling:

drop table foo;

create table foo (a int unique not null, b text);

drop function foo_ins();

create function foo_ins() returns int as '
declare
va int;
vb text;
begin
va := 1;
vb := ''xxxxx'';
INSERT INTO foo (a,b) VALUES (va,vb);
return 1;
end;
' language 'plpgsql';

select foo_ins();
foo_ins
---------
1
(1 row)

select * from foo;
a | b
---+-------
1 | xxxxx
(1 row)

Works fine here - v7.1.1 but I'm sure I'd have noticed problems in previous
versions if there'd been a fault. I'd suspect some issue with the type of
the variable or something.

What precisely is the error and is it possible to provide the table and
function definition?

- Richard Huxton

Puzzling

I am pretty sure there are some bugs in v7.1.x PL/PGSQL, they are difficult
to produce and nail down though. A few of my functions that were working
fine in v7.0.x broke completely. I am doing more testing and will submit
some bug reports...

- Andrew

"Andrew Snow" <andrew@modulus.org> writes:

I am pretty sure there are some bugs in v7.1.x PL/PGSQL, they are difficult
to produce and nail down though.

Especially if you don't even specify exactly which version you're using
;-)

Are you aware of the empty-SELECT-result-dumps-core bug in 7.1.1? That
might explain some of your difficulties ...

regards, tom lane