def reduce(self, func: Union[Callable, ReduceFunction]) -> 'DataStream':
"""
Applies a reduce transformation on the grouped data stream grouped on by the given
key position. The `ReduceFunction` will receive input values based on the key value.
Only input values with the same key will go to the same reducer.
Example:
::
>>> ds = env.from_collection([(1, 'a'), (2, 'a'), (3, 'a'), (4, 'b'])
>>> ds.key_by(lambda x: x[1]).reduce(lambda a, b: a[0] + b[0], b[1])
:param func: The ReduceFunction that is called for each element of the DataStream.
:return: The transformed DataStream.
"""
if not isinstance(func, ReduceFunction):
if callable(func):
func = ReduceFunctionWrapper(func)
else:
raise TypeError(
"The input must be a ReduceFunction or a callable function!"
)
from pyflink.fn_execution.flink_fn_execution_pb2 import UserDefinedDataStreamFunction
func_name = "m_reduce_" + str(func)
j_python_data_stream_scalar_function_operator, j_output_type_info = \
self._get_java_python_function_operator(func,
None,
func_name,
UserDefinedDataStreamFunction.REDUCE)
return DataStream(
self._j_data_stream.transform(
"Keyed Reduce", j_output_type_info,
j_python_data_stream_scalar_function_operator))
def __init__(self,
name: str,
reduce_function,
type_info: TypeInformation):
"""
Constructor of the ReducingStateDescriptor.
:param name: The name of the state.
:param reduce_function: The ReduceFunction used to aggregate the state.
:param type_info: The type of the values in the state.
"""
super(ReducingStateDescriptor, self).__init__(name, type_info)
from pyflink.datastream.functions import ReduceFunction, ReduceFunctionWrapper
if not isinstance(reduce_function, ReduceFunction):
if callable(reduce_function):
reduce_function = ReduceFunctionWrapper(reduce_function) # type: ignore
else:
raise TypeError("The input must be a ReduceFunction or a callable function!")
self._reduce_function = reduce_function
def __init__(self, name: str, reduce_function, type_info: TypeInformation):
"""
Constructor of the ReducingStateDescriptor.
:param name: The name of the state.
:param reduce_function: The ReduceFunction used to aggregate the state.
:param type_info: The type of the values in the state.
"""
super(ReducingStateDescriptor, self).__init__(name, type_info)
from pyflink.datastream.functions import ReduceFunction, ReduceFunctionWrapper
if not isinstance(reduce_function, ReduceFunction):
if callable(reduce_function):
reduce_function = ReduceFunctionWrapper(
reduce_function) # type: ignore
else:
raise TypeError(
"The input must be a ReduceFunction or a callable function!"
)
if not isinstance(type_info, PickledBytesTypeInfo):
raise ValueError(
"The type information of the state could only be PickledBytesTypeInfo "
"(created via Types.PICKLED_BYTE_ARRAY()) currently, got %s" %
type(type_info))
self._reduce_function = reduce_function