On Mon, Jul 15, 2013 at 5:56 AM, Ming Lai <[email protected]> wrote:
> I know how elog works. elog only show the status, but it does not allow me
> to execute another query when the current query fails because one of the
> invalid column was specified.
Hrm? Im not sure what you mean. If you elog(ERROR) outside of eval the
current transaction will be aborted. Thats why I suggested doing
elog(INFO) instead. The below example works fine for me. Perhaps you
can highlight exactly what you think it broken so I can understand?
=> begin;
BEGIN
=> create table a_table (a_column int);
CREATE TABLE
=> CREATE OR REPLACE FUNCTION foo() RETURNS text as $$
my $sql = "";
my $status = "";
my $r = "";
$sql = 'SELECT non_exist_column from a_table limit 1';
eval { spi_exec_query($sql);};
if ($@) {
$status = 'invalid: '.$@;
my $rv = spi_exec_query('SELECT true as col;');
return "$status\nQuery after error: ".$rv->{rows}[0]{'col'};
} else {
$status = 'valid';
}
return $status;
$$ LANGUAGE plperl;
CREATE FUNCTION
=> select foo();
foo
──────────────────────────────────────────────────────────────
invalid: column "non_exist_column" does not exist at line 6.↵
↵
Query after error: t
(1 row)
=> select true;
bool
──────
t
(1 row)
=> commit;
COMMIT
--
Sent via pgsql-bugs mailing list ([email protected])
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-bugs
