sqlalchemy.exc.StatementError with datetime field #1007
First Check
Commit to Help
Example Codefrom sqlmodel import SQLModel, Field, Session, create_engine
from datetime import datetime
class MyModel(SQLModel, table=True):
rowid: int | None = Field(default=None, primary_key=True)
timestamp: datetime
# Create an instance with a float (Unix timestamp)
instance = MyModel(timestamp=1625097600.0)
# Check the type and value
print(type(instance.timestamp)) # <class 'float'>
engine = create_engine("sqlite:////tmp/database.db")
SQLModel.metadata.create_all(engine)
with Session(engine) as session:
session.add(instance)
session.commit()Description
This throws the following error: Operating SystemLinux Operating System DetailsArch Linux SQLModel Version0.0.19 Python VersionPython 3.12.3 Additional ContextNo response |
Answered by
TheTechromancer
Jul 12, 2024
Replies: 2 comments 5 replies
|
While initializing the instance, you provided a float to the timestamp field. You need to pass a datetime object. timestamp = 1625097600.0 This is because the type of the timestamp field is datetime. Hence it must be provided a datetime object |
3 replies
|
Pydantic documentation states it will accept float:
If you remove from sqlmodel import SQLModel, Field, create_engine
from datetime import datetime
class MyModel(SQLModel):
rowid: int | None = Field(default=None, primary_key=True)
timestamp: datetime
# Create an instance with a float (Unix timestamp)
instance = MyModel(timestamp=1625097600.0)
# Check the type and value
print(type(instance.timestamp)) # <class 'datetime'> |
|
Based on the error message that you are getting, it seems that the error comes from Sqlalchemy's side. Reading the traceback suggests that the root cause of the error is related to Sqlite |
|
@20-buck-spin thank you for trying but my question is for someone familiar with the internals of SQLmodel. |
|
Found the answer in the source code. To finalize the pydantic validation, I needed to call from sqlmodel import SQLModel, Field, Session, create_engine
from datetime import datetime
class MyModel(SQLModel, table=True):
rowid: int | None = Field(default=None, primary_key=True)
timestamp: datetime
# Create an instance with a float (Unix timestamp)
instance = MyModel(timestamp=1625097600.0)
validated_instance = MyModel.model_validate(instance) # <----- THIS LINE FIXES IT
# Check the type and value
print(type(instance.timestamp)) # <class 'float'>
print(type(validated_instance.timestamp)) # <class 'datetime.datetime'>
engine = create_engine("sqlite:////tmp/database.db")
SQLModel.metadata.create_all(engine)
with Session(engine) as session:
session.add(validated_instance)
session.commit() |
2 replies
|
I'm still curious why pydantic validation isn't automatically performed when |
|
Does anyone have an answer for this ^^^? Ran into this issue today in FastAPI bump: @tiangolo edit: |
Answer selected by
TheTechromancer
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Found the answer in the source code. To finalize the pydantic validation, I needed to call
model_validate():