Class Unicode

lib/sqlalchemy/sql/sqltypes.py:307–347

A variable length Unicode string type. The :class:`.Unicode` type is a :class:`.String` subclass that assumes input and output strings that may contain non-ASCII characters, and for some backends implies an underlying column type that is explicitly supporting of non-ASCII data, such

class Unicode(String):
    """A variable length Unicode string type.

    The :class:`.Unicode` type is a :class:`.String` subclass that assumes
    input and output strings that may contain non-ASCII characters, and for
    some backends implies an underlying column type that is explicitly
    supporting of non-ASCII data, such as ``NVARCHAR`` on Oracle Database and
    SQL Server.  This will impact the output of ``CREATE TABLE`` statements and
    ``CAST`` functions at the dialect level.

    The character encoding used by the :class:`.Unicode` type that is used to
    transmit and receive data to the database is usually determined by the
    DBAPI itself. All modern DBAPIs accommodate non-ASCII strings but may have
    different methods of managing database encodings; if necessary, this
    encoding should be configured as detailed in the notes for the target DBAPI
    in the :ref:`dialect_toplevel` section.

    In modern SQLAlchemy, use of the :class:`.Unicode` datatype does not
    imply any encoding/decoding behavior within SQLAlchemy itself.  In Python
    3, all string objects are inherently Unicode capable, and SQLAlchemy
    does not produce bytestring objects nor does it accommodate a DBAPI that
    does not return Python Unicode objects in result sets for string values.

    .. warning:: Some database backends, particularly SQL Server with pyodbc,
       are known to have undesirable behaviors regarding data that is noted
       as being of ``NVARCHAR`` type as opposed to ``VARCHAR``, including
       datatype mismatch errors and non-use of indexes.  See the section
       on :meth:`.DialectEvents.do_setinputsizes` for background on working
       around unicode character issues for backends like SQL Server with
       pyodbc as well as cx_Oracle.

    .. seealso::

        :class:`.UnicodeText` - unlengthed textual counterpart
        to :class:`.Unicode`.

        :meth:`.DialectEvents.do_setinputsizes`

    """

    __visit_name__ = "unicode"


class UnicodeText(Text):