when i call "add" or "set" on a model does it do anything on the db side ?

It should not be a problem to update two tables at the same time As I understand it, the reason for the error that you show is the following: 1) enum value was assigned to Scale.source. Pony see it as an enum value (but doesn't actually support working with enums) 2) After the object was stored to the database, the value was saved as a string 3) Then object was read from the database again, in the same transaction. Pony sees the value from the database and panics, because the string value that was read from database does not exactly match enum value that was initially assigned to the attribute So, the error has nothing with updating two tables at the same time

oh, ok

seems logical

I will change my Enums then, thanks!]

i am new to pony, I don't get the diferencce between flush and commit

flush saves all unsaved objects to database, but does not commit transaction commit performs flush and then commits transaction Usually it is not necessary to call flush or commit manually: * At the end of db_session commit will be executed automatically * Before each select Pony perform flush implicitly to be sure that select take into account all changes which were made to objects Usually the only reason to perform flush manually is to get object id value if it is necessary. If id value is generated automatically you need to save object in order to see actual id value

when i call "add" or "set" on a model does it do anything on the db side ?

No, it doesn't. All changes are postponed until next flush. But it happens automatically before each select

So, if you set the same object several times at row, all changes will be combined together to be sent as a single UPDATE

brillant

when i run Token.select(lambda t: t.token == bearer and datetime.now() <= t.ts + timedelta(seconds=t.ttl))

i got NotImplementedError: timedelta(seconds=t.ttl)

I can't use this in PonyOrm? Only static?

Try t.ts >= datetime.now() - timedelta(seconds=t.ttl))

same result :)

using python pony orm with postgres(aws rds)... using it to execute raw sql ... so i created a wrapper class around pony classes to initialize database object

self.db = Database() self.db.bind(provider="postgres", user=self.username, password=self.password, host=self.hostname, database=self.database)

every time a method is called to execute raw sql query(below line), a new connection is made. when does connection pooling kick in

self.db.execute(query, query_args)

or is there a way to set connection pooling params

How can I return an object to json format?

obj.to_json()

I overwrite to_json to serialize some data type like date time from to_dict result. It also permits tu use rapidjson. I might remember Alexander told us that native to_json should not be used

Oh yes you’re right Jim, so to_dict() to get a dict and then serialize with any json library

Oh yes you’re right Jim, so to_dict() to get a dict and then serialize with any json library

so, should i use this?

hi,guys,i desgined an ER Dirgarm https://editor.ponyorm.com/user/mlzboy/zzz/designer I have a many to many relationship between Student and Exam through Score Entity,but generate the model seems loop reference,occur error could not find xxx class

I have an entity with 12 attributes, when I do a query and call to_dict method to the result object, it return me a dict with 11 attributes! it takes away one, and that one is precisely an attribute of type LongStr, does that have something to do with it?

this is my original object

and this is the resulting dictionary

I really need help here

I really need help here

docs.ponyorm.org to_dict “By default this method doesn’t include collections (to-many relationships) and lazy attributes. “

to_dict(only=None, exclude=None, with_collections=False, with_lazy=False, related_objects=False)

https://docs.ponyorm.org/api_reference.html?highlight=to_dict#Entity.to_dict

By default lazy is set to True for LongStr and LongUnicode and to False for all other types.

By default lazy is set to True for LongStr and LongUnicode and to False for all other types.

I really understand now, so the only i need is set with_lazy = True, very thanks you

I really understand now, so the only i need is set with_lazy = True, very thanks you

👍

pardon me if not completely obvious.. but is there a way to query the physical sqlite filepath of a db?

https://stackoverflow.com/questions/13912731/python-sqlite3-get-sqlite-connection-path

use db.get_connection() to get the conn object from pony

You can get filename that Pony uses to connect to the database as db.provider.pool.filename

You can get filename that Pony uses to connect to the database as db.provider.pool.filename

#todo #docs

fantastic info, ty

Is there way way to have the minimap not take away 80% of the property window?

what device are you using? what is the screen resolution?

Standart 13'' macbook, half of the screen.

1280px x 800px so the window is 640px x 800px

oh, I see we didn't consider such use case currently there is no way to adjust the prop window width

more like x 747px with the menubar and stuff.

oh, I see we didn't consider such use case currently there is no way to adjust the prop window width

Also I realized the value in the Index field is often staying when switching between properties.

Also I realized the value in the Index field is often staying when switching between properties.

thanks for reporting! #editor #bug #todo

And about the whole "Did-it-save" thing: How about a tooltip over the save button, saying something like "saved 23 seconds ago"? Something like Moment.js with Tooltip.js could help out there.

And about the whole "Did-it-save" thing: How about a tooltip over the save button, saying something like "saved 23 seconds ago"? Something like Moment.js with Tooltip.js could help out there.

#editor #todo

Luckydonald, I agree that it is better to mark "virtual" attributes such as collections by cursive

^ maybe as well?

#editor #todo

that was something I requested some time ago, but it probably just got forgotten.

Luckydonald, I agree that it is better to mark "virtual" attributes such as collections by cursive

#editor #todo

Thankies!

that was something I requested some time ago, but it probably just got forgotten.

thank you for your suggestions

I have more of those if you like.

Like the ability to actually use the "Comments" field in the SQL output. CREATE TABLE `table` ( `id` INTEGER PRIMARY KEY AUTO_INCREMENT COMMENT "bla blah blah"

Like the ability to actually use the "Comments" field in the SQL output. CREATE TABLE `table` ( `id` INTEGER PRIMARY KEY AUTO_INCREMENT COMMENT "bla blah blah"

👍 #editor #todo

And a pure json output format. Because then I could create my own templates for when forced to use more ugly languages like PHP.

Alternatively a way to upload templates or such to Ponyorm.

And a pure json output format. Because then I could create my own templates for when forced to use more ugly languages like PHP.

Also that one would be really great if there is a import button, too. I imagine it could add all the tables from the json and just overwrite existing ones.

And a pure json output format. Because then I could create my own templates for when forced to use more ugly languages like PHP.

For what? A diagram?

For what? A diagram?

If possible the whole thing, yes.

For those folks who have to take stuff under version control even if downloading and uploading is quite cumbersome. Still better as a models.txt containing a link to a snapshot.

For those folks who have to take stuff under version control even if downloading and uploading is quite cumbersome. Still better as a models.txt containing a link to a snapshot.

*Coughs in guilt*

integration with git, both the python model and json model description, correct?

integration with git, both the python model and json model description, correct?

Yeah. That would be one helpful thing.

Something went wrong: Invalid discriminator type: None

Something went wrong: Invalid discriminator type: None

That field is not storing anything when switching properties/entities in the editor.

Another idea, not as important: Having a button & dropdown to move a column between classes (especially between inheranced classes) If that functionallity is done, maybe even drag and drop could be implemented in a second step.

#Bug: When Cloning a diagram to a new name, e.g. export and that one is already existent it creates export_1 instead, however it still loads export not export_1.

Another idea, not as important: Having a button & dropdown to move a column between classes (especially between inheranced classes) If that functionallity is done, maybe even drag and drop could be implemented in a second step.

right I also would like to have this feature for myself) #editor #todo

#Bug: When Cloning a diagram to a new name, e.g. export and that one is already existent it creates export_1 instead, however it still loads export not export_1.

#editor #bug

The reason I discovered that bug is that I want to only export part of the diagram to SVG/PNG so I have to duplicate it and delete all the entities I don't want

@metaprogrammer spam

Either an admin, or telegram itself deleted that user.

Who is in charge of this group, or would be, if the creator account would not be deleted?

I.e. who to contact best about applying as admin and stuff?

At this moment it is better to contact Алексей about this stuff