Alexander

You can use lambdas too, for more complex queries: Activite.get(lambda p: p.matiere == x and p.famille == y and p.something > z)

Alexey

Jim I actually like lambdas better. How about you guys?

stsouko

Sometimes classic kwargs is useful

Jim

Since I use pony I progressively stooped to use lambdas. At the beginning I was using lambdas and generator expressions. I was always mixing everything and writing bad syntax.. So I now use select for almost everything (more "pythonic", gen exp/comprehesion is a well known syntax, more granularity to select only some fields). About the get,as can see I was a bit lost since I was using it only for for primarykeys before. Looking at it now, If possible I will use kwargs I think : shorter, dict unpacking. Thank you everyone for this short reminder

Alexey

Since I use pony I progressively stooped to use lambdas. At the beginning I was using lambdas and generator expressions. I was always mixing everything and writing bad syntax.. So I now use select for almost everything (more "pythonic", gen exp/comprehesion is a well known syntax, more granularity to select only some fields). About the get,as can see I was a bit lost since I was using it only for for primarykeys before. Looking at it now, If possible I will use kwargs I think : shorter, dict unpacking. Thank you everyone for this short reminder

Great, thanks for this info! 🌟🌟🌟

Xavier

Hi, I have a quesetion related with memory usage. The select() returns an iterator ? If I do list(select()) I use more memory?

Alexander

select() returns query object. Once query is executed it has all objects in memory.

Xavier

so there is no diference?

Alexander

Yes. How would you save the memory?

Xavier

yes

Xavier

because it's a big query

Alexander

Fetching objects one by one will take huge amount of time

Jim

maybe fetching on a batch mode ?

Jim

there is page() function in pony may it could be usefull for it (i never used it)

Alexander

As I remember page is a combination of limit and offset

Matthew

Even if you use pagination the pony cache will expand as you load more objects. You would need to empty the cache.

Alexander

At this moment Pony does not have a batch mode, because all loaded objects stays in memory until the end of db_session It is hard to say which objects can be garbage collected PonyORM uses IdentityMap pattern where object loaded during db_session are stored inside identity map and interconnected via relationships. In other ORMs which implement Active Record pattern it is easier to garbage collect objects, because ORM itself does not hold any references to previously loaded objects. But on the other hand you will have multiple copies of the same object, it it was loaded several times. Maybe in the future we can invent some way to determine which objects can be garbage collected

Lucky

At this moment Pony does not have a batch mode, because all loaded objects stays in memory until the end of db_session It is hard to say which objects can be garbage collected PonyORM uses IdentityMap pattern where object loaded during db_session are stored inside identity map and interconnected via relationships. In other ORMs which implement Active Record pattern it is easier to garbage collect objects, because ORM itself does not hold any references to previously loaded objects. But on the other hand you will have multiple copies of the same object, it it was loaded several times. Maybe in the future we can invent some way to determine which objects can be garbage collected

I think a generall clear all would be helpful though.

Alexander

It would be equivalent to commit() rollback() Note that you can’t use previously loaded objects after this

Lucky

It would be equivalent to commit() rollback() Note that you can’t use previously loaded objects after this

TiL

Abuelazo

hola, alguien asignar el atributo time ejemplo hrs = Required(Time)

Jim

hrs = Required(datetime.datetime)

Abuelazo

hrs = Required(datetime.datetime)

hrs_trab = Required(datetime.datetime) AttributeError: type object 'datetime.datetime' has no attribute 'datetime

Jim

thre is from datetime import datetime before ? only Required(datetime) then

Luis

Or import datetime and use datetime.datetime

Abuelazo

sigue dando el mismo resultado

Jim

show us your code

Abuelazo

https://pastebin.com/rip6Nbrx

Abuelazo

class Registro_actividades

Jim

You import datetime beforr line 6

Abuelazo

hrs_trab = Required(datetime.datetime) AttributeError: type object 'datetime.datetime' has no attribute 'datetime'

Alexander

As you import looks like from datetime import datetime you need to write hrs_trab = Required(datetime)

Abuelazo

solucionado muchas gracias 👍🏼

csa3d

Is it possible to store uint64 data without conversion

Alexander

attr = Required(int, size=64)

Alexander

should work

csa3d

Awesome sauce

csa3d

Ty

Matthew

with db_session: x = X.select().first() x.y += 1 commit() rollback() Will this get the object, increment its field, commit the transaction, then remove the object from the in-memory cache?

Alexander

Yes, and you can also add strict=True option to db_session to be sure objects which are interlinked by relationships are disconnected from each other, it should make garbage collector work easier

Matthew

Great! thank you. Hopefully this will reduce the memory usage of some background workers which load a lot of pony objects

Matthew

Is there any way to tell python to free memory from the heap back to the system after dropping a bunch of objects?

Matthew

Google didn’t show anything

Alexander

you can do import gc; gc.collect() but with db_session(strict=True) it is probably not necessary

Matthew

Is it a problem if I have two nested strict db_sessions?

Alexander

I think no, inner will be ignored

Matthew

thank you 🙂

Rick

hi! how can I get n random elements from a table?

Alexander

What is the primary key of the table, is it just id?

Rick

What is the primary key of the table, is it just id?

yeah

Alexander

MyEntity.select_random(n)

Rick

but I have another column (category) and I want to filter out some of these

Alexander

hmm, then:

Alexander

MyEntity.select(lambda obj: obj.category == x).random(n)

Rick

thank you very much! you're the best

Rick

MyEntity.select(lambda obj: obj.category == x).random(n)

I have this error: AttributeError: module 'parser' has no attribute 'suite'

Alexander

Can you show full traceback?

Rick

Alexander

Ok, give me a few minutes...

Rick

Sure, don't worry :P

Alexander

Do you use Python 3.8?

Rick

yeah

Alexander

It looks strange, because function parser.suite presents in Python 3.8 documentation https://docs.python.org/3/library/parser.html#parser.suite Python 3.8.0 (default, Nov 6 2019, 16:00:02) [MSC v.1916 64 bit (AMD64)] :: Anaconda, Inc. on win32 Type "help", "copyright", "credits" or "license" for more information. >>> import parser >>> parser.suite <built-in function suite> >>>

Alexander

Everything works good on my machine. Can you do some minimal script which reproduces the problem?

Rick

It looks strange, because function parser.suite presents in Python 3.8 documentation https://docs.python.org/3/library/parser.html#parser.suite Python 3.8.0 (default, Nov 6 2019, 16:00:02) [MSC v.1916 64 bit (AMD64)] :: Anaconda, Inc. on win32 Type "help", "copyright", "credits" or "license" for more information. >>> import parser >>> parser.suite <built-in function suite> >>>

Rick

Everything works good on my machine. Can you do some minimal script which reproduces the problem?

Will do! Thanks.

Rick

Everything works good on my machine. Can you do some minimal script which reproduces the problem?

https://privatebin.net/?08f0a4d3310e05fc#ETN1nq9X6TtwsecgKJj5H2ZRvK178RNQM6iRwCNNyX39 edit: traceback: https://privatebin.net/?8ce2687d50654b3c#7axg1GvWGyAL7iH2kvFqU8WCMa5tLo1Fzv8jgBbaktrn

Rick

@metaprogrammer my bad!! I had a file in the root folder named parser.py

Rick

I renamed it and now it works

Alexander

Cool

Rick

I'm very sorry!

Alexander

No problem :)

Anna

thank you very much! you're the best

I must say -- much respect for the Rick&Morty avatar!! 🔥🔥🔥

Jorge

Good morning! i have an issue with a little API using aiohttp + ponyorm

Jorge

I don't know if is a memory leak or cache

Jorge

The API is a POST operation inserting 2 entities and returing a property

Jorge

every time the api gets a request the python process up 0.1 MB

Jorge

On startup, I instantiate the database object, then, per request aiohttp call an async custom function handler and this handler call a function with the db_session decorator

Jorge

the async function handler