:bug: Fix enum type checks ordering in `get_sqlalchemy_type` by pierrecdn · Pull Request #442 · fastapi/sqlmodel

pierrecdn · 2022-09-08T22:11:49Z

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type.

e.g.:

class Foo(Enum) will work but won't allow a proper OpenAPI spec. to be generated.
class Foo(str, Enum) will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors).

Changing ordering of checks to ensure that Enum subclasses are tested first.

[1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing
[2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type. e.g.: * `class Foo(Enum)` will work but won't allow a proper OpenAPI spec. to be generated. * `class Foo(str, Enum)` will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors). Changing ordering of checks to ensure that Enum subclasses are tested first. [1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing [2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

github-actions · 2022-09-08T22:13:17Z

📝 Docs preview for commit a6d9568 at: https://631a68e284fe1d20404a6714--sqlmodel.netlify.app

codecov · 2022-09-08T22:14:25Z

Codecov Report

Merging #442 (a6d9568) into main (75ce455) will not change coverage.
The diff coverage is 100.00%.

@@ Coverage Diff @@## main #442 +/- ##
=======================================
Coverage 97.76% 97.76% =======================================
Files 187 187 Lines 6268 6268 =======================================
Hits 6128 6128 Misses 140 140

Impacted Files	Coverage Δ
sqlmodel/main.py	`84.92% <100.00%> (ø)`

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

pierrecdn · 2022-09-08T22:19:23Z

Hmm, seeing that I'm not the only one to observe this:

Enum types on inherited models don't correctly create type in Postgres #164 (comment)
Add option to pass kwargs to SQLAlchemy types #366 (with a different approach though)

@tiangolo any thoughts?

daniil-berg · 2022-09-09T07:19:26Z

This makes sense. It may need to be rebased with PR #425.

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

pierrecdn · 2022-09-09T07:56:05Z

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

Indeed, just seen you proposal (#436), sounds really nice.
The PR is quite large though, so it might take longer to get merged.

tiangolo · 2023-10-23T09:23:28Z

Thank you!

As I can't push to this branch, I included your commits in a new branch in a new PR here: #669

Given that, I'll close this one. Thanks for the work! 🙇

pierrecdn · 2022-09-08T22:11:49Z

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type.

e.g.:

class Foo(Enum) will work but won't allow a proper OpenAPI spec. to be generated.
class Foo(str, Enum) will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors).

Changing ordering of checks to ensure that Enum subclasses are tested first.

[1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing
[2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type. e.g.: * `class Foo(Enum)` will work but won't allow a proper OpenAPI spec. to be generated. * `class Foo(str, Enum)` will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors). Changing ordering of checks to ensure that Enum subclasses are tested first. [1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing [2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

github-actions · 2022-09-08T22:13:17Z

📝 Docs preview for commit a6d9568 at: https://631a68e284fe1d20404a6714--sqlmodel.netlify.app

codecov · 2022-09-08T22:14:25Z

Codecov Report

Merging #442 (a6d9568) into main (75ce455) will not change coverage.
The diff coverage is 100.00%.

@@ Coverage Diff @@## main #442 +/- ##
=======================================
Coverage 97.76% 97.76% =======================================
Files 187 187 Lines 6268 6268 =======================================
Hits 6128 6128 Misses 140 140

Impacted Files	Coverage Δ
sqlmodel/main.py	`84.92% <100.00%> (ø)`

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

pierrecdn · 2022-09-08T22:19:23Z

Hmm, seeing that I'm not the only one to observe this:

Enum types on inherited models don't correctly create type in Postgres #164 (comment)
Add option to pass kwargs to SQLAlchemy types #366 (with a different approach though)

@tiangolo any thoughts?

daniil-berg · 2022-09-09T07:19:26Z

This makes sense. It may need to be rebased with PR #425.

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

pierrecdn · 2022-09-09T07:56:05Z

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

Indeed, just seen you proposal (#436), sounds really nice.
The PR is quite large though, so it might take longer to get merged.

tiangolo · 2023-10-23T09:23:28Z

Thank you!

As I can't push to this branch, I included your commits in a new branch in a new PR here: #669

Given that, I'll close this one. Thanks for the work! 🙇

pierrecdn · 2022-09-08T22:11:49Z

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type.

e.g.:

class Foo(Enum) will work but won't allow a proper OpenAPI spec. to be generated.
class Foo(str, Enum) will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors).

Changing ordering of checks to ensure that Enum subclasses are tested first.

[1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing
[2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type. e.g.: * `class Foo(Enum)` will work but won't allow a proper OpenAPI spec. to be generated. * `class Foo(str, Enum)` will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors). Changing ordering of checks to ensure that Enum subclasses are tested first. [1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing [2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

github-actions · 2022-09-08T22:13:17Z

📝 Docs preview for commit a6d9568 at: https://631a68e284fe1d20404a6714--sqlmodel.netlify.app

codecov · 2022-09-08T22:14:25Z

Codecov Report

Merging #442 (a6d9568) into main (75ce455) will not change coverage.
The diff coverage is 100.00%.

@@ Coverage Diff @@## main #442 +/- ##
=======================================
Coverage 97.76% 97.76% =======================================
Files 187 187 Lines 6268 6268 =======================================
Hits 6128 6128 Misses 140 140

Impacted Files	Coverage Δ
sqlmodel/main.py	`84.92% <100.00%> (ø)`

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

pierrecdn · 2022-09-08T22:19:23Z

Hmm, seeing that I'm not the only one to observe this:

Enum types on inherited models don't correctly create type in Postgres #164 (comment)
Add option to pass kwargs to SQLAlchemy types #366 (with a different approach though)

@tiangolo any thoughts?

daniil-berg · 2022-09-09T07:19:26Z

This makes sense. It may need to be rebased with PR #425.

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

pierrecdn · 2022-09-09T07:56:05Z

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

Indeed, just seen you proposal (#436), sounds really nice.
The PR is quite large though, so it might take longer to get merged.

tiangolo · 2023-10-23T09:23:28Z

Thank you!

As I can't push to this branch, I included your commits in a new branch in a new PR here: #669

Given that, I'll close this one. Thanks for the work! 🙇

pierrecdn · 2022-09-08T22:11:49Z

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type.

e.g.:

class Foo(Enum) will work but won't allow a proper OpenAPI spec. to be generated.
class Foo(str, Enum) will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors).

Changing ordering of checks to ensure that Enum subclasses are tested first.

[1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing
[2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type. e.g.: * `class Foo(Enum)` will work but won't allow a proper OpenAPI spec. to be generated. * `class Foo(str, Enum)` will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors). Changing ordering of checks to ensure that Enum subclasses are tested first. [1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing [2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

github-actions · 2022-09-08T22:13:17Z

📝 Docs preview for commit a6d9568 at: https://631a68e284fe1d20404a6714--sqlmodel.netlify.app

codecov · 2022-09-08T22:14:25Z

Codecov Report

Merging #442 (a6d9568) into main (75ce455) will not change coverage.
The diff coverage is 100.00%.

@@ Coverage Diff @@## main #442 +/- ##
=======================================
Coverage 97.76% 97.76% =======================================
Files 187 187 Lines 6268 6268 =======================================
Hits 6128 6128 Misses 140 140

Impacted Files	Coverage Δ
sqlmodel/main.py	`84.92% <100.00%> (ø)`

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

pierrecdn · 2022-09-08T22:19:23Z

Hmm, seeing that I'm not the only one to observe this:

Enum types on inherited models don't correctly create type in Postgres #164 (comment)
Add option to pass kwargs to SQLAlchemy types #366 (with a different approach though)

@tiangolo any thoughts?

daniil-berg · 2022-09-09T07:19:26Z

This makes sense. It may need to be rebased with PR #425.

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

pierrecdn · 2022-09-09T07:56:05Z

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

Indeed, just seen you proposal (#436), sounds really nice.
The PR is quite large though, so it might take longer to get merged.

tiangolo · 2023-10-23T09:23:28Z

Thank you!

As I can't push to this branch, I included your commits in a new branch in a new PR here: #669

Given that, I'll close this one. Thanks for the work! 🙇

pierrecdn · 2022-09-08T22:11:49Z

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type.

e.g.:

class Foo(Enum) will work but won't allow a proper OpenAPI spec. to be generated.
class Foo(str, Enum) will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors).

Changing ordering of checks to ensure that Enum subclasses are tested first.

[1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing
[2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type. e.g.: * `class Foo(Enum)` will work but won't allow a proper OpenAPI spec. to be generated. * `class Foo(str, Enum)` will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors). Changing ordering of checks to ensure that Enum subclasses are tested first. [1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing [2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

github-actions · 2022-09-08T22:13:17Z

📝 Docs preview for commit a6d9568 at: https://631a68e284fe1d20404a6714--sqlmodel.netlify.app

codecov · 2022-09-08T22:14:25Z

Codecov Report

Merging #442 (a6d9568) into main (75ce455) will not change coverage.
The diff coverage is 100.00%.

@@ Coverage Diff @@## main #442 +/- ##
=======================================
Coverage 97.76% 97.76% =======================================
Files 187 187 Lines 6268 6268 =======================================
Hits 6128 6128 Misses 140 140

Impacted Files	Coverage Δ
sqlmodel/main.py	`84.92% <100.00%> (ø)`

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

pierrecdn · 2022-09-08T22:19:23Z

Hmm, seeing that I'm not the only one to observe this:

Enum types on inherited models don't correctly create type in Postgres #164 (comment)
Add option to pass kwargs to SQLAlchemy types #366 (with a different approach though)

@tiangolo any thoughts?

daniil-berg · 2022-09-09T07:19:26Z

This makes sense. It may need to be rebased with PR #425.

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

pierrecdn · 2022-09-09T07:56:05Z

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

Indeed, just seen you proposal (#436), sounds really nice.
The PR is quite large though, so it might take longer to get merged.

tiangolo · 2023-10-23T09:23:28Z

Thank you!

As I can't push to this branch, I included your commits in a new branch in a new PR here: #669

Given that, I'll close this one. Thanks for the work! 🙇

pierrecdn · 2022-09-08T22:11:49Z

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type.

e.g.:

class Foo(Enum) will work but won't allow a proper OpenAPI spec. to be generated.
class Foo(str, Enum) will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors).

Changing ordering of checks to ensure that Enum subclasses are tested first.

[1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing
[2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type. e.g.: * `class Foo(Enum)` will work but won't allow a proper OpenAPI spec. to be generated. * `class Foo(str, Enum)` will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors). Changing ordering of checks to ensure that Enum subclasses are tested first. [1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing [2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

github-actions · 2022-09-08T22:13:17Z

📝 Docs preview for commit a6d9568 at: https://631a68e284fe1d20404a6714--sqlmodel.netlify.app

codecov · 2022-09-08T22:14:25Z

Codecov Report

Merging #442 (a6d9568) into main (75ce455) will not change coverage.
The diff coverage is 100.00%.

@@ Coverage Diff @@## main #442 +/- ##
=======================================
Coverage 97.76% 97.76% =======================================
Files 187 187 Lines 6268 6268 =======================================
Hits 6128 6128 Misses 140 140

Impacted Files	Coverage Δ
sqlmodel/main.py	`84.92% <100.00%> (ø)`

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

pierrecdn · 2022-09-08T22:19:23Z

Hmm, seeing that I'm not the only one to observe this:

Enum types on inherited models don't correctly create type in Postgres #164 (comment)
Add option to pass kwargs to SQLAlchemy types #366 (with a different approach though)

@tiangolo any thoughts?

daniil-berg · 2022-09-09T07:19:26Z

This makes sense. It may need to be rebased with PR #425.

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

pierrecdn · 2022-09-09T07:56:05Z

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

Indeed, just seen you proposal (#436), sounds really nice.
The PR is quite large though, so it might take longer to get merged.

tiangolo · 2023-10-23T09:23:28Z

Thank you!

As I can't push to this branch, I included your commits in a new branch in a new PR here: #669

Given that, I'll close this one. Thanks for the work! 🙇

pierrecdn · 2022-09-08T22:11:49Z

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type.

e.g.:

class Foo(Enum) will work but won't allow a proper OpenAPI spec. to be generated.
class Foo(str, Enum) will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors).

Changing ordering of checks to ensure that Enum subclasses are tested first.

[1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing
[2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type. e.g.: * `class Foo(Enum)` will work but won't allow a proper OpenAPI spec. to be generated. * `class Foo(str, Enum)` will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors). Changing ordering of checks to ensure that Enum subclasses are tested first. [1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing [2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

github-actions · 2022-09-08T22:13:17Z

📝 Docs preview for commit a6d9568 at: https://631a68e284fe1d20404a6714--sqlmodel.netlify.app

codecov · 2022-09-08T22:14:25Z

Codecov Report

Merging #442 (a6d9568) into main (75ce455) will not change coverage.
The diff coverage is 100.00%.

@@ Coverage Diff @@## main #442 +/- ##
=======================================
Coverage 97.76% 97.76% =======================================
Files 187 187 Lines 6268 6268 =======================================
Hits 6128 6128 Misses 140 140

Impacted Files	Coverage Δ
sqlmodel/main.py	`84.92% <100.00%> (ø)`

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

pierrecdn · 2022-09-08T22:19:23Z

Hmm, seeing that I'm not the only one to observe this:

Enum types on inherited models don't correctly create type in Postgres #164 (comment)
Add option to pass kwargs to SQLAlchemy types #366 (with a different approach though)

@tiangolo any thoughts?

daniil-berg · 2022-09-09T07:19:26Z

This makes sense. It may need to be rebased with PR #425.

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

pierrecdn · 2022-09-09T07:56:05Z

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

Indeed, just seen you proposal (#436), sounds really nice.
The PR is quite large though, so it might take longer to get merged.

tiangolo · 2023-10-23T09:23:28Z

Thank you!

As I can't push to this branch, I included your commits in a new branch in a new PR here: #669

Given that, I'll close this one. Thanks for the work! 🙇

pierrecdn · 2022-09-08T22:11:49Z

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type.

e.g.:

class Foo(Enum) will work but won't allow a proper OpenAPI spec. to be generated.
class Foo(str, Enum) will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors).

Changing ordering of checks to ensure that Enum subclasses are tested first.

[1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing
[2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

When enum subclassing/mix-in[1] is used, especially to ensure compatibility and produce correct OpenAPI specifications with FastAPI[2], SQLModel on its side doesn't map to the correct SQLAlchemy type. e.g.: * `class Foo(Enum)` will work but won't allow a proper OpenAPI spec. to be generated. * `class Foo(str, Enum)` will fix the previous situation, but prevents SQLModel to read and map the object from the DB (due to validation errors). Changing ordering of checks to ensure that Enum subclasses are tested first. [1] https://docs.python.org/3/library/enum.html#restricted-enum-subclassing [2] https://fastapi.tiangolo.com/sq/tutorial/path-params/#create-an-enum-class

github-actions · 2022-09-08T22:13:17Z

📝 Docs preview for commit a6d9568 at: https://631a68e284fe1d20404a6714--sqlmodel.netlify.app

codecov · 2022-09-08T22:14:25Z

Codecov Report

Merging #442 (a6d9568) into main (75ce455) will not change coverage.
The diff coverage is 100.00%.

@@ Coverage Diff @@## main #442 +/- ##
=======================================
Coverage 97.76% 97.76% =======================================
Files 187 187 Lines 6268 6268 =======================================
Hits 6128 6128 Misses 140 140

Impacted Files	Coverage Δ
sqlmodel/main.py	`84.92% <100.00%> (ø)`

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

pierrecdn · 2022-09-08T22:19:23Z

Hmm, seeing that I'm not the only one to observe this:

Enum types on inherited models don't correctly create type in Postgres #164 (comment)
Add option to pass kwargs to SQLAlchemy types #366 (with a different approach though)

@tiangolo any thoughts?

daniil-berg · 2022-09-09T07:19:26Z

This makes sense. It may need to be rebased with PR #425.

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

pierrecdn · 2022-09-09T07:56:05Z

If PR #436 is merged, you'll be able to specify the SQL type explicitly whenever you want. Then PR #366 will become obsolete. It's still worth maintaining/optimizing get_sqlachemy_type though to cover the simple and straightforward cases, so that you don't need to specify the SQL type.

Indeed, just seen you proposal (#436), sounds really nice.
The PR is quite large though, so it might take longer to get merged.

tiangolo · 2023-10-23T09:23:28Z

Thank you!

As I can't push to this branch, I included your commits in a new branch in a new PR here: #669

Given that, I'll close this one. Thanks for the work! 🙇

tiangolo added the investigate label Nov 28, 2022

tepelbaum mentioned this pull request Nov 29, 2022
Enum types on inherited models don't correctly create type in Postgres #164
Closed
8 tasks

tiangolo added the bug Something isn't working label Oct 22, 2023

tiangolo mentioned this pull request Oct 23, 2023
🐛 Fix enum type checks ordering in get_sqlalchemy_type#669
Merged

tiangolo closed this Oct 23, 2023