[docs]@dataclass(init=False)classColumn(CastMixin):r"""A column represents metadata information for a column in a Kumo :class:`~kumoai.graph.Table`. Columns can be created independent of a table, or can be fetched from a table with the :meth:`~kumoai.graph.Table.column` method. .. code-block:: python import kumoai # Fetch a column from a `kumoai.Table`: table = kumoai.Table(...) column = table.column("col_name") column = table["col_name"] # equivalent to the above. # Edit a column's data type: print("Existing dtype: ", column.dtype) column.dtype = "int" # Edit a column's semantic type: print("Existing stype: ", column.stype) column.stype = "ID" Args: name: The name of this column. stype: The semantic type of this column. Semantic types can be specified as strings: the list of possible semantic types is located at :class:`~kumoai.Stype`. dtype: The data type of this column. Data types can be specified as strings: the list of possible data types is located at :class:`~kumoai.Dtype`. timestamp_format: If this column represents a timestamp, the format that the timestamp should be parsed in. The format can either be a :class:`~kumoapi.table.TimestampUnit` for integer columns or a string with a format identifier described `here <https://spark.apache.org/docs/latest/sql-ref-datetime-pattern.html>`__ for a SaaS Kumo deployment and `here <https://docs.snowflake.com/en/sql-reference/date-time-input-output#about-the-elements-used-in-input-and-output-formats>`__ for a Snowpark Container Services Kumo deployment. If left empty, will be intelligently inferred by Kumo. """# noqa: E501name:strstype:Optional[Stype]=Nonedtype:Optional[Dtype]=Nonetimestamp_format:Optional[Union[str,TimestampUnit]]=None