[SPARK-51919][PYTHON] Allow overwriting statically registered Python Data Source #50716
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@allisonwang-db @HyukjinKwon please take a look |
wengh
commented
Apr 25, 2025
| - During Data Source resolution, built-in and Scala/Java Data Sources take precedence over Python Data Sources with the same name; to explicitly use a Python Data Source, make sure its name does not conflict with the other Data Sources. | ||
| - During Data Source resolution, built-in and Scala/Java Data Sources take precedence over Python Data Sources with the same name; to explicitly use a Python Data Source, make sure its name does not conflict with the other non-Python Data Sources. | ||
| - It is allowed to register multiple Python Data Sources with the same name. Later registrations will overwrite earlier ones. | ||
| - To automatically register a data source, export it as ``DefaultSource`` in a top level module with name prefix ``pyspark_``. See `pyspark_huggingface <https://github.com/huggingface/pyspark_huggingface>`_ for an example. |
There was a problem hiding this comment.
Should we mention the DefaultSource feature which was previously undocumented?
wengh
changed the title
HyukjinKwon
approved these changes
Apr 27, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
Why are the changes needed?
Static registration is a bit obscure and doesn't always work as expected (e.g. when the module providing DefaultSource is installed after
lookup_data_sourcesalready ran).So in practice users (or LLM agents) often want to explicitly register the data source even if it is provided as a DefaultSource.
Raising an error in this case interrupts the workflow, making LLM agents spend extra tokens regenerating the same code but without registration.
This change also makes the behavior consistent with user data source registration which are already allowed to overwrite previous user registrations.
Does this PR introduce any user-facing change?
Yes. Previously, registering a Python Data Source with the same name as a statically registered one would throw an error. With this change, it will overwrite the static registration.
How was this patch tested?
Added a test in
PythonDataSourceSuite.scalato verify that static sources can be overwritten correctly.Was this patch authored or co-authored using generative AI tooling?
No