Best way to insert multiple rows with asyncpg
asyncpg
provides the executemany
method to insert many rows.
statement = """INSERT INTO films (code,
title,
did,
date_prod,
kind) VALUES($1, $2, $3, $4, $5);"""
await connection.executemany(statement, values)
If you need to use RETURNING
as you later mentioned to return the inserted ids, this answer is the way to go.
If you need to use the RETURNING
clause to obtain the ids back, then the following is the most efficient way of inserting multiple values:
res = await conn.fetch('''
INSERT INTO films (code, title, did, date_prod, kind)
(SELECT
r.code, r.title, r.did, r.date_prod, r.kind
FROM
unnest($1::films[]) as r
)
RETURNING id
''', [
(None, 'B6717', 'Tampopo', 110, '1985-02-10', 'Comedy'),
(None, 'HG120', 'The Dinner Game', 140, None, 'Comedy')
])
Note that the records you pass as input must correspond to the shape of the table: PostgreSQL does not support arbitrary records as input, so you must use a known record type. Simply pass the columns you are not inserting as None
and don't include them in the SELECT
return list. This method also doesn't allow you to rely on DEFAULT
, you must specify each inserted value explicitly.
Another way to insert many rows at once (assuming you don't need the inserted IDs) is to use the copy_records_to_table
method.
data = [
("row", 1, "some data"),
("row", 2, "more data"),
]
await conn.copy_records_to_table('mytable', records=data)