HI Adrian,
I had a typo in the email:
INSERT INTO my_table VALUES(my_table.*);
was actually
INSERT INTO my_table VALUES(my_var.*);
So I meant to insert the variable I had in memory (dict representing a row), not the rows from the table..
> On 02/06/2015, at 01:44, Adrian Klaver <adrian.klaver@aklaver.com> wrote:
>
> On 06/01/2015 07:42 AM, Filipe Pina wrote:
>> Thanks for the reply anyway, it's a pity though, it'd be useful..
>>
>> Another bump I've found along the pl/python road: insert ROWTYPE in table..
>> Maybe you have some hint on that? :)
>>
>> So, in PLPGSQL I can:
>>
>> DECLARE
>> my_var my_table;
>> BEGIN
>> my_var.col1 := 'asd';
>> INSERT INTO my_table VALUES(my_table.*);
>> END;
>>
>> How would I do something like that in pl/python?
>>
>> First, how to declare a ROW-TYPE variable, as they're all python mappings?
>>
>> my_var = { 'col1': 'asd' } enough? it'd would miss all the other columns...
>>
>> Second, how to insert it?
>>
>> plpy.prepare and .execute say they don't support composite types, so I
>> cannot simply pass
>>
>> pl = plpy.prepare('INSERT INTO core_customer VALUES ($1)', ['my_table'])
>>
>> Any workarounds for this? (meaning I wouldn't have to specify any
>> columns in the insert statement)
>
> http://www.postgresql.org/docs/9.4/interactive/sql-insert.html
>
> pl = plpy.prepare('INSERT INTO core_table SELECT * FROM my_table')
>
>>
>> Thanks
>>
>> On Sex, Mai 29, 2015 at 2:00 , Peter Eisentraut <peter_e@gmx.net> wrote:
>>> On 5/18/15 10:52 AM, Filipe Pina wrote:
>>>
>>> But one of the functions I need to create needs to accept an array
>>> of records.
>>>
>>> PL/Python doesn't support that. Some more code needs to be written to
>>> support that. You did everything correctly. I don't know of a good
>>> workaround.
>
>
> --
> Adrian Klaver
> adrian.klaver@aklaver.com