[
https://issues.apache.org/jira/browse/BEAM-12932?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Marek Suscak updated BEAM-12932:
--------------------------------
Description:
[Unlike|https://github.com/apache/beam/blob/a92965969fda3d7be5b65996d1a9a5b7dd373119/sdks/python/apache_beam/io/fileio.py#L476-L483]
[FileIO
module|https://beam.apache.org/releases/pydoc/2.32.0/apache_beam.io.fileio.html],
the
[FileBasedSink|https://beam.apache.org/releases/pydoc/2.32.0/apache_beam.io.filebasedsink.html]
module [doesn't
respect|https://github.com/apache/beam/blob/2c619c81082839e054f16efee9311b9f74b6e436/sdks/python/apache_beam/io/filebasedsink.py#L164-L177]
the
_[temp_location|https://beam.apache.org/releases/pydoc/2.32.0/_modules/apache_beam/options/pipeline_options.html]_
pipeline option. This isn't expected and documented properly in the respective
module's documentation.
I do understand that the _temp_location_ [is technically a pipeline option
that's only required by the
DataflowRunner|https://github.com/apache/beam/blob/cf8e08f51bc9b1b25f6ee8007dfcc0fa04322e15/sdks/python/apache_beam/options/pipeline_options.py#L626-L631]
but I think it might be worth it to pluck it out of the DataflowRunner and
make it be a global pipeline option.
In addition, I couldn't info on what is the difference between
[WriteToFiles|https://beam.apache.org/releases/pydoc/2.32.0/apache_beam.io.fileio.html]
and
[WriteToText|https://beam.apache.org/releases/pydoc/2.32.0/apache_beam.io.textio.html]
transforms. It seems that the former uses the FileBasedSink and thus correctly
writes temporary files to the temp_location when specified but the latter
doesn't do that.
was:
[Unlike|https://github.com/apache/beam/blob/a92965969fda3d7be5b65996d1a9a5b7dd373119/sdks/python/apache_beam/io/fileio.py#L476-L483]
[FileIO
module|https://beam.apache.org/releases/pydoc/2.32.0/apache_beam.io.fileio.html],
the
[FileBasedSink|https://beam.apache.org/releases/pydoc/2.32.0/apache_beam.io.filebasedsink.html]
module [doesn't
respect|https://github.com/apache/beam/blob/2c619c81082839e054f16efee9311b9f74b6e436/sdks/python/apache_beam/io/filebasedsink.py#L164-L177]
the
_[temp_location|https://beam.apache.org/releases/pydoc/2.32.0/_modules/apache_beam/options/pipeline_options.html]_
pipeline option. This isn't expected and documented properly in the respective
module's documentation
In addition, I couldn't info on what is the difference between
[WriteToFiles|https://beam.apache.org/releases/pydoc/2.32.0/apache_beam.io.fileio.html]
and
[WriteToText|https://beam.apache.org/releases/pydoc/2.32.0/apache_beam.io.textio.html]
transforms. It seems that the former uses the FileBasedSink and thus correctly
writes temporary files to the temp_location when specified but the latter
doesn't do that.
> File based sink doesn't respect temp_location pipeline option
> -------------------------------------------------------------
>
> Key: BEAM-12932
> URL: https://issues.apache.org/jira/browse/BEAM-12932
> Project: Beam
> Issue Type: Bug
> Components: io-py-files
> Affects Versions: 2.32.0
> Reporter: Marek Suscak
> Priority: P4
> Labels: documentation, file-component, python
>
> [Unlike|https://github.com/apache/beam/blob/a92965969fda3d7be5b65996d1a9a5b7dd373119/sdks/python/apache_beam/io/fileio.py#L476-L483]
> [FileIO
> module|https://beam.apache.org/releases/pydoc/2.32.0/apache_beam.io.fileio.html],
> the
> [FileBasedSink|https://beam.apache.org/releases/pydoc/2.32.0/apache_beam.io.filebasedsink.html]
> module [doesn't
> respect|https://github.com/apache/beam/blob/2c619c81082839e054f16efee9311b9f74b6e436/sdks/python/apache_beam/io/filebasedsink.py#L164-L177]
> the
> _[temp_location|https://beam.apache.org/releases/pydoc/2.32.0/_modules/apache_beam/options/pipeline_options.html]_
> pipeline option. This isn't expected and documented properly in the
> respective module's documentation.
> I do understand that the _temp_location_ [is technically a pipeline option
> that's only required by the
> DataflowRunner|https://github.com/apache/beam/blob/cf8e08f51bc9b1b25f6ee8007dfcc0fa04322e15/sdks/python/apache_beam/options/pipeline_options.py#L626-L631]
> but I think it might be worth it to pluck it out of the DataflowRunner and
> make it be a global pipeline option.
> In addition, I couldn't info on what is the difference between
> [WriteToFiles|https://beam.apache.org/releases/pydoc/2.32.0/apache_beam.io.fileio.html]
> and
> [WriteToText|https://beam.apache.org/releases/pydoc/2.32.0/apache_beam.io.textio.html]
> transforms. It seems that the former uses the FileBasedSink and thus
> correctly writes temporary files to the temp_location when specified but the
> latter doesn't do that.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
