Re: does the executemany() limitation with result sets apply to INSERT..RETURNING ?

On Mon, Mar 2, 2009 at 5:32 PM, Michael Bayer <[email protected]> wrote:
>
> On Mar 2, 2009, at 3:09 AM, Federico Di Gregorio wrote:
>
>> Il giorno dom, 01/03/2009 alle 15.07 -0500, Michael Bayer ha scritto:
>>>
>>> - executemany() now return the numer of total INSERTed or UPDATEd
>>>   rows. Note that, as it has always been, executemany() should not
>>>   be used to execute multiple SELECT statements and while it will
>>>   execute the statements without any problem, it will return the
>>>   wrong value.
>>>
>>> This appears to be the case as well for INSERT...RETURNING as in the
>>> example below.
>>
>> This has always been the case with any statement returning tuples.
>> executemany() *is not* for gathering multiple result sets.
>
> well the only difference here is that an INSERT..RETURNING would in theory
> return just one result set even for many statement executions, since each
> insert only affects a single row.

Not quite true.  The following query will give a result set containing
three rows:

    insert into foo (id) values (1), (2), (3) returning id;


> but don't worry, im not asking for this, just confirming the behavior.

Note that psycopg doesn't parse the SQL statements: it just sends them
through to PostgreSQL.  The response from PostgreSQL then indicates
whether the statement produced a result set.

So you shouldn't really expect psycopg to exhibit special behaviour
for particular statements.

James.