SQLAlchemy can act as an Object-relational Map (ORM). This enables you to treat records of a database table as instantiable records. To leverage this functionality, you must reflect the underlying metadata in one of the following ways.

Note: The following examples employ SQLAlchemy 1.4.

Modeling Data Using a Mapping Class

Use "sqlalchemy.ext.declarative.declarative_base" to declare a mapping class for the table you wish to model in the ORM. A known table in the data model is modeled either partially or completely, as shown in the following example:

from sqlalchemy.ext.declarative import declarative_base
Base = declarative_base()
class [CData].[Human_Resources].Workers(Base):
	__tablename__ = "[CData].[Human_Resources].Workers"
	Id = Column(String, primary_key=True)
	Worker_Reference_WID = Column(String)
	Legal_Name_Last_Name = Column(String)

Automatically Reflecting Metadata

Rather than mapping tables manually, SQLAlchemy can discover the metadata for one or more tables automatically. To accomplish this across the entire data model, use automap_base:

from sqlalchemy import MetaData
from sqlalchemy.ext.automap import automap_base
meta = MetaData()
abase = automap_base(metadata=meta)
abase.prepare(autoload_with=engine)
[CData].[Human_Resources].Workers = abase.classes.[CData].[Human_Resources].Workers

You can also reflect a single table with an inspector. When reflecting this way, providing a list of specific columns to map is optional:

from sqlalchemy import MetaData, Table
from sqlalchemy import inspect
meta = MetaData()
insp = inspect(engine)
[CData].[Human_Resources].Workers_table = Table("[CData].[Human_Resources].Workers", meta)
insp.reflect_table([CData].[Human_Resources].Workers_table, ["Id","Legal_Name_Last_Name"])