TheNeuralBit commented on a change in pull request #11766: URL: https://github.com/apache/beam/pull/11766#discussion_r433537957
########## File path: sdks/python/apache_beam/dataframe/partitionings.py ########## @@ -0,0 +1,133 @@ +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +from __future__ import absolute_import + +from typing import Any +from typing import Iterable +from typing import TypeVar + +import pandas as pd + +Frame = TypeVar('Frame', bound=pd.core.generic.NDFrame) + + +class Partitioning(object): + """A class representing a (consistent) partitioning of dataframe objects. + """ + def is_subpartition_of(self, other): + # type: (Partitioning) -> bool + + """Returns whether self is a sub-partition of other. + + Specifically, returns whether something partitioned by self is necissarily + also partitioned by other. + """ + raise NotImplementedError + + def partition_fn(self, df): + # type: (Frame) -> Iterable[Tuple[Any, Frame]] + + """A callable that actually performs the partitioning of a Frame df. + + This will be invoked via a FlatMap in conjunction with a GroupKey to + achieve the desired partitioning. + """ + raise NotImplementedError + + +class Index(Partitioning): + """A partitioning by index (either fully or partially). + + If the set of "levels" of the index to consider is not specified, the entire + index is used. + + These form a partial order, given by + + Nothing() < Index([i]) < Index([i, j]) < ... < Index() < Singleton() Review comment: This ordering is determined by `is_subpartition_of` correct? I wonder if there's a way to clearly say that in this docstring? ########## File path: sdks/python/apache_beam/dataframe/frames_test.py ########## @@ -23,6 +23,7 @@ from apache_beam.dataframe import expressions from apache_beam.dataframe import frame_base +from apache_beam.dataframe import frames # pylint: disable=unused-import Review comment: What is this for? ########## File path: sdks/python/apache_beam/dataframe/partitionings.py ########## @@ -0,0 +1,133 @@ +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +from __future__ import absolute_import + +from typing import Any +from typing import Iterable +from typing import TypeVar + +import pandas as pd + +Frame = TypeVar('Frame', bound=pd.core.generic.NDFrame) + + +class Partitioning(object): + """A class representing a (consistent) partitioning of dataframe objects. + """ + def is_subpartition_of(self, other): + # type: (Partitioning) -> bool + + """Returns whether self is a sub-partition of other. + + Specifically, returns whether something partitioned by self is necissarily + also partitioned by other. + """ + raise NotImplementedError + + def partition_fn(self, df): + # type: (Frame) -> Iterable[Tuple[Any, Frame]] + + """A callable that actually performs the partitioning of a Frame df. + + This will be invoked via a FlatMap in conjunction with a GroupKey to + achieve the desired partitioning. + """ + raise NotImplementedError + + +class Index(Partitioning): + """A partitioning by index (either fully or partially). + + If the set of "levels" of the index to consider is not specified, the entire + index is used. + + These form a partial order, given by + + Nothing() < Index([i]) < Index([i, j]) < ... < Index() < Singleton() + """ + + _INDEX_PARTITIONS = 100 Review comment: Previously this was 10 right (in `partitioned_by_index`)? Assuming this intentional, but I just wanted to double-check its not a typo. ########## File path: sdks/python/apache_beam/dataframe/partitionings.py ########## @@ -0,0 +1,133 @@ +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +from __future__ import absolute_import + +from typing import Any +from typing import Iterable +from typing import TypeVar + +import pandas as pd + +Frame = TypeVar('Frame', bound=pd.core.generic.NDFrame) + + +class Partitioning(object): + """A class representing a (consistent) partitioning of dataframe objects. + """ + def is_subpartition_of(self, other): Review comment: nit: I think I'd prefer `is_subpartitioning_of` ########## File path: sdks/python/apache_beam/dataframe/expressions.py ########## @@ -85,16 +87,10 @@ def evaluate_at(self, session): # type: (Session) -> T """Returns the result of self with the bindings given in session.""" raise NotImplementedError(type(self)) - def requires_partition_by_index(self): # type: () -> bool - """Whether this expression requires its argument(s) to be partitioned - by index.""" - # TODO: It might be necessary to support partitioning by part of the index, - # for some args, which would require returning more than a boolean here. + def requires_partition_by(self): # type: () -> Partitioning raise NotImplementedError(type(self)) - def preserves_partition_by_index(self): # type: () -> bool - """Whether the result of this expression will be partitioned by index - whenever all of its inputs are partitioned by index.""" + def preserves_partition_by(self): # type: () -> Partitioning Review comment: The meaning of this function is a little confusing now since it implies some connection to the input partitioning, but it also has it's own partitioning. Would renaming it to `outputs_..` or `produces_..` still be accurate, or is the output partitioning actually a function of both "preserves" and the input? I also think we should consider changing `.._partition_by` to `.._partitioning` for clarity. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
