And than you can find a dog by primary key: dog = Dog[dog_id] And find an owner of this dog: owner = dog.owner

Yeah i know xD

But i have like. Its a town with dogs that "have" many homes

And i want to get every person that have that dog as a pet

Like, those dogs choose where to eat, spend the time and sleep

Then probably the relation should be many-to-many (two Set attributes). But anyway, if any human of this city is required to have a dog, then select(h for h in Human if h.pet == dog) looks correct. But the error says that the value of a dog variable is not a single dog, but a set of dogs. You can print the variable and see its value.

hmmm ok

i get

Dog[3]

Then it is strange that the error says that you are comparing Animal with a Set of Dog. I don't see any Set here

me neither 😆

but that's exactly what i have and the query i'm trying to do

and that's what i get u.u

(the error)

Can you provide full exception traceback? It may contain some hints

Dont know how xD

The error `<class 'pony.orm.sqltranslation.IncomparableTypesError'> ("Incomparable types 'Animal' and 'Set of Dog' in expression: h.pet == dog",)` came with some traceback

well, i don't know what maybe i've done wrong before

but now that works

sorry for the trouble

noow i have another question

if i have a function that returns an User object, and User inheritates from db.Entity

if the .. scope that called the function that returned that User object it's wrapped in db_entity it will get that is the same user as in the table?

or i should "load" it again?

Yes, within the same db_session it will be the same object if primary key value is the same

thanks :D

Annnnnnnnnother silly question(sorry, i guess i don't have too much experience u.u)

Hmm well basically i would ask "what could i do if the database is locked"

i was thinking of just using @retry at best

or try to make a buffer

but to me that seems a little odd while using pony 😆

Can you describe the scenario when you get the database locked?

having 3 processes that access the same database at the same time

In SQLite only one process can modify database at a time. After the process issues an INSERT, UPDATE or DELETE command, the database is in locked state. If another process tries to work with the database at this moment, it will be suspended until the first process finishes its transaction. SQLite has timeout (which defaults to 5 seconds), after which the suspended process got exception "The database is locked". In order to avoid that exception, you need to complete transaction in the first process before the timeout. In Pony a transaction can be finished in following ways: - implicitly, upon exit from db_session scope - explicitly, using commit() or rollback() Your process that saved changes to database should perform commit() or rollback() after some portion of data was saved, or exit from db_session in order to have implicit commit. It should not stay within db_session indefinitely without performin commit or rollback

mm

to use the retry option

only can used on @db_session?

like..

i cant do with db_session(retry=3): #code

?

Yes, this is Python restriction for context managers, their code cannot be automatically executed several times

wokey ty 😆

while db_session.error(retry=3): # codeeverything is possible in python ;)

No, here is it with db_session.retry(3) as retry: while retry: # code

So you mean that we can add such extension to API

NO! :)

another round of "Rozen's stupid questions!" yey!

if i have db1 and db1 class hello(db1.Entity): blablablal

can i have like... class wololo(db2.Entity): helloes = Set("Hello") ?

hey nope, all related entities should belong to the same database

yes you can afaik

but if they are not related, then you can

right, i didn't notice they were related

hmm wokey

and if i have... a relational table in one of the databases?

well, actually i have no idea what i'm talking about 😆

haha)

Why do you want to split tables between several databases? It is better to keep related tables together, because then the database can enforce foreign key constraints in order to be sure that all used ids have corresponding rows. It is possible to keep tables in different databases, but then you are vulnerable to data consistency issues and need to process joins manually.

it was just an idea 😆

By default try to store everything in a single database, it will be simpler

yes yes 😆 i was and idea beause i wanted to avoid to have lots of database locking 😆

i'm having tons of database writing at the same time 😆

i have retry=30 and everything but it crashes anyway 😆

That means that db_session from which you perform an update stays open for too long. Ideally, all transactions which perform writing to the database should be very short. Maybe you start db_session, perform an update and then start some unrelated actions, like, sending e-mail via SMTP. In that case the database remains locked until the email is sent, and it may take a long time. It is better to perform such long unrelated operations outside of db_session

mmmmmmmmm

i'll see

Smaller transactions is always better. If you have to retry 30 times, think about why you wanted to use a db again :p

ajamm

This works? with db_session: hoomans = select(h for h in Human) for h in hoomans: print(h.name)

(or it's even necesary to use de db_session for just a select?)

You should wrap all interactions with objects in a single db_session: with db_session: hoomans = select(...) for h in hoomans: print(h.name) An attribute access like h.name potentially can lead to SQL query, and SQL queries are possible inside db_session only.

ok ok, got it 😆

i'm not quite sure how generators work so maaaybeee it worked (?)

mm

but if i do hoomans = (h.name for h in Human) ?

generators are lazy, the real operation will be performed later. You need to use list comprehension to eval result inside db_session: with db_session: hoomans = select(...) names = [h.name for h in hooman] Then you can use names outside of db_session

right

thanks :D

Helleau!

hello :D