Using Indexing When Performing JOINs on Part of a Composite Key

I have a table in my database (hosted on MariaDB) that looks like this (I have given the definition in SQLAlchemy), class Address(Base): __tablename__ = 'address' __table_args__ = {'schema': DB_NAME} referenceID = sqlalchemy.Column(sqlalchemy.String(length=25), primary_key=True) referenceTable = sqlalchemy.Column(sqlalchemy.String(length=25), primary_key=True) sourceReferenceID = sqlalchemy.Column(sqlalchemy.String(length=25), primary_key=True) addressType = sqlalchemy.Column(sqlalchemy.String(length=25), primary_key=True) zip = sqlalchemy.Column(sqlalchemy.String(length=10)) streetAddress = sqlalchemy.Column(sqlalchemy.String(length=50)) city = sqlalchemy.Column(sqlalchemy.String(length=25)) fullAddress = sqlalchemy.Column(sqlalchemy.String(length=200)) country = sqlalchemy.Column(sqlalchemy.String(length=25)) validFrom = sqlalchemy.Column(sqlalchemy.Date) validTo = sqlalchemy.Column(sqlalchemy.Date) Now, it can be seen that my table has a composite key which includes several attributes. As far as I know, a clustered index will be created out of the primary key of a table. I am not completely aware of how this works when the table consists of a composite key. An explanation on this would be appreciated. Further, I am going to want to perform several JOINs on this table, particularly using the referenceID and sourceReferenceID attributes of the table. To do this efficiently, do I need to create indexes (non-clustered) on these attributes separately using index=True?