Repository navigation
[FLINK-41028] Support data lake table properties in the Fluss pipeline sink - #4584
Conversation
…e sink Co-Authored-By: Codex <noreply@openai.com> AI-Model: gpt-6-astra AI-Contributed/Feature: 116/116 AI-Contributed/UT: 136/136
| admin.getTableInfo(new TablePath(DATABASE_NAME, "table1")) | ||
| .get() | ||
| .toTableDescriptor(); | ||
| assertThat(tableDescriptor.getCustomProperties()) |
There was a problem hiding this comment.
Overall, this PR looks good to me. This test verifies that the options are stored in the Fluss TableDescriptor as custom properties, but it does not enable lake storage or check the resulting lake table. Would it be possible to add an end-to-end test that creates a lake-enabled Fluss table through the pipeline sink and asserts that an option such as properties.paimon.file.format: orc is applied to the actual Paimon table? That would cover the full path from sink configuration to the lake table.
There was a problem hiding this comment.
Thanks for the suggestion. I agree that an end-to-end test would be valuable, but it would require introducing Paimon dependencies and substantial additional test setup. Could we track this in a separate follow-up task for lake integration testing, keeping this PR focused on passing custom properties through the pipeline sink?
AI-Model: gpt-6-astra
AI-Contributed/Feature: 116/116
AI-Contributed/UT: 136/136
What is the purpose of this pull request?
Briefly describe the problem this PR fixes or the feature it introduces. Reference Flink JIRA ticket when possible.
Brief change log
List the key changes in this PR.
Verifying this change
This change is a trivial rework / code cleanup without any test coverage.
(or)
This change added tests and can be verified as follows:
...*...*Documentation
Was generative AI tooling used to co-author this PR?