It is signed in that case

@luckydonald I think that you already created table with integer column type. Try to re-create table so it will use bigint type instead

As you just started working on your project, I think it should be easy to just drop old tables and recreate them

Right now we are working on a migration tool which allow to automatically generate ALTER TABLE... commands to reflect changes in entity defnitions

OK. I thought int, size=64 would already create the fitting database type? https://www.postgresql.org/docs/9.1/static/datatype-numeric.html Or is it just used in the internal range checks? Anyway, Required(int, sql_type='bigint', size=64, unsigned=False) indeed works, thank you.

int, size=64 should be enough, but if table was already created Pony does not attempt to modify it when you change model definition in Python

Okey, I probably forgot to drop it. Much thanks.

int, size=64 should be enough, but if table was already created Pony does not attempt to modify it when you change model definition in Python

Yeah, can confirm now.

Actually I can scroll it, but the minimap is in the way

fixed can you check if it works for you?

Maybe you should include that in that paragraph, that depending on the database there is something needed.

added this info to https://docs.ponyorm.com/firststeps.html#installing

Welcome!

File "/usr/local/lib/python3.5/site-packages/pony/orm/core.py", line 627, in _exec_raw_sql assert locals is None AssertionError

It seems I have to specify globals={}, too? Maybe a more verbose error? And/Or warn in https://docs.ponyorm.com/queries.html#using-the-select-by-sql-and-get-by-sql-methods ?

Anyone getting an Idea what that means?

For reference, my data model, if that helps: https://editor.ponyorm.com/user/luckydonald/Tags

Trying to get a sticker from the StickerMessage table, with given message_id and chat_id(==user_id)

> It seems I have to specify globals={}, too? Maybe a more verbose error? In standard Python functions (like eval) if you pass just one dict it called globals: https://docs.python.org/3.6/library/functions.html#eval But such functions typically accept positional arguments only. Probably we need to add handling of situations, where locals passed without globals

> Anyone getting an Idea what that means? What Python version do you use? Try to replace [...] with (...) inside select

> It seems I have to specify globals={}, too? Maybe a more verbose error? In standard Python functions (like eval) if you pass just one dict it called globals: https://docs.python.org/3.6/library/functions.html#eval But such functions typically accept positional arguments only. Probably we need to add handling of situations, where locals passed without globals

Maybe I should use globals instead of locals, don't actually know what they do. But letting the orm pull variables from the context feels unsecure to me. Also my IDE is complaining about unused variables :D

> Anyone getting an Idea what that means? What Python version do you use? Try to replace [...] with (...) inside select

python3.5. I'll try that

Works, awesome. Thanks

How can I use the OFFSET in the pythonic style? I try to write the SQL in the first line as python query below

.limit(50, offset=offset)

That was unexpected to be in limit

Is there a Github page with Wiki, maybe we could start making a list mapping SQL statements to ponyorm?

If offset is called separately it will alter the meaning of previous `limit`call

I see

Is there a Github page with Wiki, maybe we could start making a list mapping SQL statements to ponyorm?

Just created one https://github.com/ponyorm/pony/wiki/SQL-statements-represented-in-PonyORM Can you check if you can edit the content?

Added that.

Added that.

very cool!

How can I check the lower case version of a text field? select( for t in Tag if lower(t.string) == lower("TestText") )

Just as in Python: t.string.lower()

Oh, right. My bad.

We plan to add special upsert method. Right now you can write something like that: user = User.get(id=user_id) if user is not None: user.name = user_name else: user = User(id=user_id, name=user_name)

How can I make sure I won't get problems with multible instances? I.e. the common variable +1 problem?

what do you mean by multiple instances?

I have multible instances of a telegram bot. Now you can spam that bot, (E.g. by forwarding messages to it) And it need to maintain a state, to get this question-answer style input working. Multible Instances would mean, some could be a little bit faster, and crate the User row already, so hat the else part failes, etc:

it should be the library user's responsibility to keep model access synchronous

if you really do have issues with multiple processes, you might want to keep a cache or something for that

to lock access while one worker is already doing it

> How can I make sure I won't get problems with multible instances? In Pony @db_session decorator has retry option. If it is set to value greater than zero, Pony will re-execute function wrapped with db_session specified number of times in case of exception. Contrived example: @db_session(retry=5) def post_message(user_id, user_name, chat_id, message_text): user = get_or_create_user(user_id, user_name) chat = get_or_create_chat(chat_id) message = Message(user=user, chat=chat, text=message_text) def get_or_create_user(user_id, user_name): user = User.get(id=user_id) if user is not None: user.name = name else: user = User(id=user_id, name=user_name) return user def get_or_create_chat(chat_id): return Chat.get(id=chat_id) or Chat(id=chat.id) In this example, if an exception ocurred during db_session execution, Pony will re-execute post_message function up to 5 times. Note that nested db_session s are ingored, only top-level db_session is taken into account.

is this a transaction?

will the changes before the exception be rewound?

entering db_session starts a transaction. If exception occurred, all changes during db_session will be rolled back. If retry is specified, Pony will start new transaction and call function again with the same arguments

ok, cool :)

Welcome!

Thanks!

Do you know of any other Python related groups on Telegram?

There are russian ones. Do you interested in them ?

uff, only speak spanish and english

On slack there are huge channels

Gitter is great if you're looking for language/library communities

Btw, what do you think of moving to Slack?

IRC not mentioned?

I use slack but wouldn't use it for something like Pony. Gitter seems more appropriate because it's almost directly related to the repo

Plus once you're there you can subscribe to many more project communities

Whats wrong with Telegram?

Nothing wrong We can see that Slack has wider use

Do you know of any other Python related groups on Telegram?

There is http://telegram.me/pytg_group for python telegram bots

I use slack but wouldn't use it for something like Pony. Gitter seems more appropriate because it's almost directly related to the repo

Why you wouldn't use it for Pony?

It's usually better for larger communities that need multiple channels to begin with

There's barely any activity in our single chatroom here, don't see why move to a "bigger room"

That's just how I see it anyway

Slack might be more convenient, no?

I guess so, for people who don't use Telegram.

I'll stay here with you guys :D

I like Telegram more :)

Hello, is this an appropriate place to ask a question about how to fix a problem I'm having with Pony?

Hello, sure

Thank you :) I've posted a StackOverflow question here: http://stackoverflow.com/questions/40871284/ . I'm being told that a value is being changed outside the current transaction, but I don't think I have multiple transactions occurring

Ok, give me a few minutes :)

Ok, I'll answer on StackOverlow

Thank you so much!

Done

Thank you for your response! That is very helpful.

I assume this problem will go away once bulk updates are implemented in the orm?

Also, you mention that volatile attributes will not be re-read from the db until the object is saved, so the value might be obsolete when read. I just want to confirm that you are talking about a potential race condition here, if two separate transactions DO, in fact, try to read and modify the object simultaneously?

Well, it may be not easy to invalidate cached attributes in Python after the bulk update was issued. We need to understand which rows were changed, and it may be not easy to do just looking at query in Python. Maybe we can invalidate entire column for all objects of the same type. In the future we'll add List attribute type which will manage columns like order_id automatically

Very cool. Thank you!

Sure!