DB_DSN in Sanic extension docs.
Hello, i have smth problem with big query set:
My service receive GraphQL request and call 3-5 method like this:
engine = self.__db_engine
slots = []
async with engine.acquire(reuse=True, reusable=True, lazy=True) as conn:
for slot_model in self.SLOT_MODEL_LIST:
query = select([Slot.id.label('base_slot_id'), Slot.agent_id, Slot.attrs_view, Slot.name, Slot.type, slot_model]).select_from(Slot.join(slot_model, slot_model.slot_id == Slot.id))
if options.agent_id:
query = query.where(Slot.agent_id == options.agent_id)
elif options.slot_id_list:
query = query.where(Slot.id.in_(options.slot_id_list))
else:
return None
async for slot in conn.iterate(query.order_by(Slot.id)):
slots.append(
slot
)
return [dto_slot_factory.from_db_model(slot) for slot in slots] #1000 - 3000 elemetsAfter the query is finished, i run gc, but the process still consumes a memory. After 20-30 query, service eating 380 mb.
My db setting:
engine = await db.set_bind(
database_url,
min_size=CONFIGURATION['DB_POOL_MIN_SIZE'], #1
max_size=CONFIGURATION['DB_POOL_MAX_SIZE'], #100
statement_cache_size=0,
)what can be done to free memory ?
Hi, I need help
from gino import Gino
db = Gino()
class MyTable(db.Model):
__tablename__ = "test"
id = db.Column(DB.Integer, primary_key=True)
name = db.Column(DB.String(60), nullable=False)
img_name = db.StringProperty()
pin_code = db.Column(DB.String(10), nullable=True)After starting I got raise AttributeError(
AttributeError: Requires "profile" JSON[B] column
Can I use StringProperty without profile field?
StringProperty is a convenient class for the string field in a JSON or JSONB column. You can surely use something other than "profile", but we'll need a JSON or JSONB to contain that.
Or rather, I think you're looking for
db.Column(db.Text()) for a string column in a db table.
class MyTable(db.Model):
_img_name = db.Column("img_name", db.Text())
@property
def img_name(self):
return self._img_name
@img_name.setter
def img_name(self, val):
self._img_name = some logic
@xmrdsc That depends by the definition of "fast" in your context. I'll say the performance of CPU resources is not the target. Coming to IO operations, it's a different story.
There are pros/cons using gevent vs asyncpg. There are a lot of discussions you can find. One related part is that using Gino now only supports psql, while gevent can "support" all other libraries.
I think it should not be too difficult to get benchmarks of both implementations.
These might be helpful: https://github.com/python-gino/gino/blob/master/docs/explanation/why.rst
python-gino/gino#420
Hi there! We use Gino for aiohttp web app, but also in the same project we have some workers and cli interfaces where synchronous db calls are made. In this synchronous places I can use SQLAlchemy core for this purposes, but I'd like to use SQLAlchemyORM to work with instances of classes created with Gino db instance. I also learned that if I change Gino db.Model to Base=declarative_base() SqlAlchemyORM works perfectly (while Gino calls loose some functionality as stated in docs).
My question is:
a) is there any way to extract SqlAlchemyORM compatible object to use in synchronous query from Gino model class (__table__ doesn't suit me as queries with it act like alchemy core not orm)
b) maybe I can use Gino itself synchronously?
c) If both are negative, maybe you know another workaround other than duplicating tables?
d) maybe I miss something conceptually important and do wrong things? :)
