>>> I understand from the code that the COPY is really executed, so the ERROR
>>> and so ROW_COUNT about the SQL should reflect that. Basically the change
>>> makes the client believe that there is an SQL error whereas the error is
>>> on the client.
>>
>> Right, but wether COPY fails because psql can't write the output,
>> possibly half-way because of a disk full condition, or because the
>> query was cancelled or the server went down, are these distinctions
>> meaningful for a script?
>
> It could if the SQL command has side effects, but probably this does not
> apply to COPY TO which cannot have.
Yes it can:
COPY (
UPDATE pgbench_branches
SET bbalance = bbalance + 1
WHERE bid <= 5
RETURNING *) TO STDOUT \g /BAD
The SQL command is executed but the backslash command fails.
--
Fabien.
