It was pointed out to me off-list that pg_dump fails if a user-defined
datatype's output routine is silly enough to emit NOTICE messages:
the rest of the table data simply disappears without any warning.
What is happening is that libpq's pqGetCopyData3() drops out of COPY_OUT
state upon seeing a N
On Fri, Jun 15, 2001 at 10:22:53AM -0400, Tom Lane wrote:
> What PG version? Do you have any open transactions that might have
> created or deleted as-yet-uncommitted tables?
I'm not sure since this has not happened on my system. Sorry, I wasn't
precise enough. It happens on the system of a co-w
Michael Meskes <[EMAIL PROTECTED]> writes:
> Could anyone please tell me what the following mean:
> NOTICE: Index pg_type_typname_index: NUMBER OF INDEX' TUPLES (124) IS NOT THE SAME
>AS HEAP' (114)
> NOTICE: Index pg_attribute_attrelid_index: NUMBER OF INDEX' TUPLES (883) IS NOT THE
>SAME AS
Could anyone please tell me what the following mean:
NOTICE: Index pg_type_typname_index: NUMBER OF INDEX' TUPLES (124) IS NOT THE SAME AS
HEAP' (114)
NOTICE: Index pg_attribute_attrelid_index: NUMBER OF INDEX' TUPLES (883) IS NOT THE
SAME AS HEAP' (422)
I've never seen them before.
Michae
