TypeError when using Enum columns with the Databricks dialect (all 2.0.x releases)

[sha16384]

Description

Any SQLAlchemy operation that involves an Enum-typed column with the Databricks dialect raises:

TypeError: String.__init__() got an unexpected keyword argument '_enums'

Steps to reproduce

 from sqlalchemy import create_engine, Column, Enum
 from sqlalchemy.orm import DeclarativeBase
 
 class Base(DeclarativeBase):
     pass
 
 class MyTable(Base):
     __tablename__ = "my_table"
     status = Column(Enum("A", "B", "C"), nullable=False)
 
 engine = create_engine("databricks://...")
 Base.metadata.create_all(engine)

Root cause

databricks-sqlalchemy registers DatabricksStringType as the colspec for sqlalchemy.String in the Databricks dialect. When SQLAlchemy adapts an Enum column (via Enum.adapt()), it internally passes _enums as a keyword argument to the target type's constructor. DatabricksStringType.__init__() does not accept **kwargs or _enums, causing the crash.

Environment

Affected versions: 2.0.1 – 2.0.9
SQLAlchemy: 2.0.x

Related issues

• Databricks dialects does not seem to support Type Json #32 - No JSON colspec, CompileError at DDL compile time
• data types: add sqlalchemy type class for the interval data type #37 - No INTERVAL colspec, Silent missing feature
• This issue - Enum colspec exists but target class rejects _enums kwarg