ok, thx!

Hi, i have two hooks - before/after update and i print the current status of where_to_post in each hook, somehow i get both the same, as you can see - each pair of two prints [one from before_update and one from after_update] is framed. i don't know why before and after they are still the same.

when i print outside the hook

i made some test, as far as i noticed, somehow before_update get called after the object has updated

@metaprogrammer i found an error, i don't know whether it's intended or not but: from pony.orm import * db = Database() class Entity(db.Entity): id = PrimaryKey(int, auto=True) num = Optional(int) db.bind('sqlite', 'b', create_db=True) db.generate_mapping(create_tables=True) en = Entity() gives error -> if entity._database_.schema is None: AttributeError: 'NoneType' object has no attribute 'schema' but when i rename the class [for any other name] it works fine. the issue is when the entity name is "Entity"

You cannot name an Entity subclass as Entity, this is a special reserved name

Ah ok. nice

In usual Python programming, when you inherit one class from another, you typically also give it a different name

i get an issues with hooks again :( from pony.orm import * db = Database() class E(db.Entity): id = PrimaryKey(int, auto=True) num = Optional(int, default=10) def before_update(self): print(self.num) def after_update(self): print(self.num) db.bind('sqlite', 'b', create_db=True) db.generate_mapping(create_tables=True) with db_session: en = E() print(en.num) en.num = 1 print(en.num) although en.num get changed, the functions after/before_update doesn't get called

Let me check it...

thanks

So: * before_update - is called before "UPDATE" SQL command execution (not directly before executing UPDATE command, but some time before it) * after_update - is called after "UPDATE" SQL command was executed (not directly after it, but some time after) * before_insert, after_insert - the same for "INSERT" command * before_delete, after_delete - the same for "DELETE" command In this specific example you create a new object en = E() It will be added to the database with "INSERT" command. If you add before_insert or after_insert hook it will be called Note that when you perform en.num = 1 it does not generate any direct SQL command. The change is accumulated into object and will be saved when performing implicit or explicit flush(). When your code exits from the db_session, it causes implicit commit(), commit causes flush(), flush causes a single INSERT, and before/after this insert the before_insert/after_insert hooks will be called

> somehow before_update get called after the object has updated If you mean that you do obj.foo = "bar" and before_update hook is called somewhere after this line, it is the correct behavior, because before_update calls before "UPDATE" sql command, not before assigning a new value to an object attribute

ok, got you. Thanks. something like def before_update(self): self._temp_change_ = self.to_dict() def after_update(self): print("before:", self._temp_change_, "after:", self.to_dict()) would be cool, so i can see the difference before/after change, as far as i understood you, i can't do that :/

You can add some function which uses internal Pony structures to get this information: def get_diff(self): result = {} for attr in self._attrs_with_bit_(self._attrs_with_columns_, self._wbits_): old_val = self._dbvals_[attr] new_val = self._vals_[attr] result[attr.name] = old_val, new_val return result Note that obj._dbvals_ stores values in form as they were read from the database, and obj._vals_ stores values in form as suitable for Python. In most cases they are the same, but for some types they may be different. For example, JSON data will be in serialized form in obj._dbvals_ (as a string), while in obj._vals_ JSON data will be represented as a TrackedDict instance (a sublcass of Python dict class)

You can add some function which uses internal Pony structures to get this information: def get_diff(self): result = {} for attr in self._attrs_with_bit_(self._attrs_with_columns_, self._wbits_): old_val = self._dbvals_[attr] new_val = self._vals_[attr] result[attr.name] = old_val, new_val return result Note that obj._dbvals_ stores values in form as they were read from the database, and obj._vals_ stores values in form as suitable for Python. In most cases they are the same, but for some types they may be different. For example, JSON data will be in serialized form in obj._dbvals_ (as a string), while in obj._vals_ JSON data will be represented as a TrackedDict instance (a sublcass of Python dict class)

Thanks!

Language: py3 Source: class A: a = lambda self: print('before:', self.__dict__) b = lambda self: print('after:', self.__dict__) def __setattr__(self, k, v): self.a() self.__dict__[k] = v self.b() a = A() a.a = 2 Result: before: {} after: {'a': 2}

i guess something like that is okay as well, thanks Alex

[with super()]

If you want to combine it with Pony entities, it may be too brittle, and doesn't catch all changes. Like class Student(db.Entity): name = Required(str) group = Optional("Group") class Group(db.Entity): number = PrimaryKey(int) students = Set("Student") group1 = Group(number=123) student1 = Student(name='John', group=group1) print(student1.group) # Group[123] group1.delete() print(student1.group) # None As objects are interrelated, sometimes attribute changes happen implicitly, due to changes in another objects

Maybe it is better not to mix additional magic to Pony entities (as they already contain pretty big amount of magic), but build a separate layer above it

ah, thanks!

currently what i have is as in the photo, but i may replace it with a layer of abstraction

Hi there, I use pony and now have a problem with the Python-Installer. The following error is displayed in the console when starting: C:\PythonDevelopment\Projects\blz_db\setup\dist\main_blz_db>main_blz_db.exe Traceback (most recent call last): File "main_blz_db.py", line 15, in <module> File "main_blz_db.py", line 9, in main File "gui\main_wnd.py", line 24, in __init__ File "common\user.py", line 77, in __init__ File "common\user.py", line 106, in __bind_db File "pony\orm\core.py", line 774, in bind File "pony\orm\core.py", line 793, in _bind File "pony\utils\utils.py", line 219, in import_module ModuleNotFoundError: No module named 'pony.orm.dbproviders' [17492] Failed to execute script main_blz_db What must be entered in "hiddenimport" for the pyinstaller script?

Hi there, I use pony and now have a problem with the Python-Installer. The following error is displayed in the console when starting: C:\PythonDevelopment\Projects\blz_db\setup\dist\main_blz_db>main_blz_db.exe Traceback (most recent call last): File "main_blz_db.py", line 15, in <module> File "main_blz_db.py", line 9, in main File "gui\main_wnd.py", line 24, in __init__ File "common\user.py", line 77, in __init__ File "common\user.py", line 106, in __bind_db File "pony\orm\core.py", line 774, in bind File "pony\orm\core.py", line 793, in _bind File "pony\utils\utils.py", line 219, in import_module ModuleNotFoundError: No module named 'pony.orm.dbproviders' [17492] Failed to execute script main_blz_db What must be entered in "hiddenimport" for the pyinstaller script?

You need to explicitly import database module from pony.orm.dbproviders which you use in your program, like pony.orm.dbproviders.sqlite

Good Morning, I now explicitly imported the module. First in my program: from pony.orm.dbproviders.sqlite import * Then in the Pyinstaller script: class PYENV(IntEnum): LOCAL = 0, VIRTUAL = auto() def run_installer(env: PYENV = PYENV.VIRTUAL): installer_path: str = '' if env == PYENV.VIRTUAL: installer_path = "C:/PythonDevelopment/Projects/blz_db/venv/Scripts/pyinstaller.exe" else: installer_path: str = "C:/PythonDevelopment/Python380_32/Scripts/pyinstaller.exe" script_path: str = "C:/PythonDevelopment/Projects/blz_db/src/main_blz_db.py" # list for commandline parameters args: List[str] = list() args.append(installer_path) # option -y clear output-dir args.append("-y") # args.append("--windowed") args.append("--onedir") args.append("--hidden-import=pony.orm.dbproviders.sqlite") # args.append("--noconsole") args.append("--distpath=C:/PythonDevelopment/Projects/blz_db/setup/dist") args.append("--workpath=C:/PythonDevelopment/Projects/blz_db/setup/build") # call pyinstaller.exe with commandline parameters args.append(script_path) subprocess.call(args) And here the error message in the console: C:\PythonDevelopment\Projects\blz_db\setup\dist\main_blz_db>main_blz_db.exe Traceback (most recent call last): File "main_blz_db.py", line 15, in <module> File "main_blz_db.py", line 9, in main File "gui\main_wnd.py", line 24, in __init__ File "common\user.py", line 78, in __init__ File "common\user.py", line 107, in __bind_db File "pony\orm\core.py", line 774, in bind File "pony\orm\core.py", line 796, in _bind File "pony\orm\dbproviders\sqlite.py", line 344, in __init__ File "pony\orm\dbapiprovider.py", line 129, in __init__ File "pony\orm\dbproviders\sqlite.py", line 455, in get_pool File "pony\utils\utils.py", line 232, in absolutize_path File "pony\utils\utils.py", line 229, in is_absolute_path TypeError: expected string or bytes-like object [11952] Failed to execute script 😭

File "common\user.py", line 107, in __bind_db It seems you pass some incorrect filename here

Not really, the _bind method is in user.py.

Not really, the _bind method is in user.py.

I don't unsertand what you mean

Check what options you pass to db.bind()

Maybe we get on wrong! My program works. But the pyinstaller does not include the pony module in the package. Here is part of the script. Everything works very well! def __bind_db(self, db_path: str = None) -> None: if db_path is None: db.bind(provider='sqlite', filename=':memory') else: db.bind(provider='sqlite', filename=db_path, create_db=True) db.generate_mapping(create_tables=True) self.__is_init = True

The last exception that you show: TypeError: expected string or bytes-like object Happens in this function, which expect filename as a string or bytes-like object: def is_absolute_path(filename): return bool(_absolute_re.match(filename)) It seems that db_path value that you pass into __bid_db is not actually a str. You can add a debug print to check what is the actual type of db_path value

The value is a string. But had, following your advice, the following explicitly imported: from pony.orm import * from pony.orm.dbproviders.sqlite import * < new ! After that, my script stopped working. I have now removed the second line again. Works again. I'll look at that with the parameter again.

> After that, my script stopped working When you run it with PyInstaller or without?

debug

and this

This is without PyInstaller, as I understand it? And the error happens when you run it with PyInstaller?

If yes, I assume that with PyInstaller you have different value for db_path

The error occurs with the package that PyInstaller has created.

But in debug you test it without PyInstaller

Can you add print(repr(db_path)) to __bind_db so you can see the result of it with PyInstaller

C:\PythonDevelopment\Projects\blz_db\setup\dist>cd main_blz_db C:\PythonDevelopment\Projects\blz_db\setup\dist\main_blz_db>main_blz_db.exe 'C:\\PythonDevelopment\\Projects\\blz_db\\setup\\dist\\main_blz_db/login.sqlite3' Traceback (most recent call last): File "pony\orm\dbapiprovider.py", line 55, in wrap_dbapi_exceptions File "pony\orm\dbapiprovider.py", line 230, in connect File "pony\orm\dbapiprovider.py", line 351, in connect File "pony\orm\dbproviders\sqlite.py", line 660, in _connect sqlite3.OperationalError: unable to open database file During handling of the above exception, another exception occurred: Traceback (most recent call last): File "main_blz_db.py", line 16, in <module> File "main_blz_db.py", line 10, in main File "gui\main_wnd.py", line 26, in __init__ File "gui\stack_pages\pg_main\pg_main.py", line 33, in __init__ File "dal_component\dal.py", line 37, in connect File "pony\orm\core.py", line 774, in bind File "pony\orm\core.py", line 796, in _bind File "pony\orm\dbproviders\sqlite.py", line 344, in __init__ File "pony\orm\dbapiprovider.py", line 130, in __init__ File "<string>", line 2, in connect File "pony\orm\dbapiprovider.py", line 72, in wrap_dbapi_exceptions pony.orm.dbapiprovider.OperationalError: unable to open database file [14152] Failed to execute script main_blz_db

You show some different traceback now. Previously the last frame of user code was File "common\user.py", line 107, in __bind_db and not it is File "dal_component\dal.py", line 37, in connect

Yes, I just changed something. I had created db_path with join (path and name). I do that now with path + / name. Join had created a list. For whatever reason.

Probably you still need to add print(repr(db_path)) to your code to see what exactly you pass to Pony as a filename

line 37 : db.bind(provider='sqlite', filename=db_path, create_db=True) I'll debug that right now!

filename is 'blz_db.sqlite3' . it's okay !

No, it is not ok. Pass an absolute path if you run it from exe

Doesn't the database have to be recreated in the current directory?

It determines absolute path relative to the python module __filename__, and I suspect that with PyInstaller you don't have correct __filename__ for each included Python source file

You're right !!!! I would never have thought of that. It works now . Many Thanks !

Sure

userlist = select (u for u in User) userlist=userlist.where(lambda u: contact (u.first_name, " ", u.last_name) = "test user")

Will this work

Basically I need to use concat in where condition

If you have a limited number of items, you can use +

userlist = select (u for u in User) userlist=userlist.where(lambda u: u.first_name+" "+ u.last_name = "test user")

Like this?

yes

It may be inefficient, as such a query cannot use indexes, even if you define index for first_name and last_name. But it should work

Hi. The link to docs in https://blog.ponyorm.org/ is broken

there is a 's' missing http://doc.ponyorm.com/

https://docs.ponyorm.org/

there is a 's' missing http://doc.ponyorm.com/

Thank you, fixed!

my entity: class Message(database.Entity): id = orm.PrimaryKey(int, default=generator) i create instance by Message() since i don't need to supply id, generator does it. but i get TypeError: The first positional argument must be lambda function or its text source. Got: 9836765217163256848074682327137247327874400116519830960375032 9836765217163256848074682327137247327874400116519830960375032 is the value that generator() returns

Technically a function isn't a lambda. Maybe default=lambda: generator() could work?

i have tried this too. doesn't work

my entity: class Message(database.Entity): id = orm.PrimaryKey(int, default=generator) i create instance by Message() since i don't need to supply id, generator does it. but i get TypeError: The first positional argument must be lambda function or its text source. Got: 9836765217163256848074682327137247327874400116519830960375032 9836765217163256848074682327137247327874400116519830960375032 is the value that generator() returns

Can I ask why would you need something like that :) ?

I don't know what is the reason for the error, but the value looks too big for int (even for int64), maybe you should use str instead of int

how about lambda: next(generator())

well, my goal is just storing things in DB but ID doesn't matter, do i just use random id and i used a large number to avoid duplicated-primary-key

how about lambda: next(generator())

it doesn't yield. it does return

Note you can use uid = PrimaryKey(UUID, default=uuid4)

just use an auto increment primary key, don't generate random IDs

Note you can use uid = PrimaryKey(UUID, default=uuid4)

oh, yes. Thanks!

or UUID as Alexander says 🙂

ok, i forgot about auto=True 🙈

misread, since I saw generator I thought it was a generator 🙂

works now

misread, since I saw generator I thought it was a generator 🙂

right, it is confusing, thanks