Fixing DatabricksNotebookOperator invalid dependency graph issue by dheerajkumar-solanki · Pull Request #48492 · apache/airflow

dheerajkumar-solanki · 2025-03-28T11:11:03Z

Fixing the DatabricksNotebookOperator invalid dependency graph issue, this comes when dependencies are defined in DatabricksWorkflowTaskGroup. Issue is child task DatabricksNotebookOperator and DatabricksTaskOperator task referencing itself. Issue in _generate_databricks_task_key function which is incorrectly generating depends_on task value.

closes: #47983

^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in airflow-core/newsfragments.

boring-cyborg · 2025-03-28T11:11:08Z

Congratulations on your first Pull Request and welcome to the Apache Airflow community! If you have any issues or are unsure about any anything please check our Contributors' Guide (https://github.com/apache/airflow/blob/main/contributing-docs/README.rst)
Here are some useful points:

Pay attention to the quality of your code (ruff, mypy and type annotations). Our pre-commits will help you with that.
In case of a new feature add useful documentation (in docstrings or in docs/ directory). Adding a new operator? Check this short guide Consider adding an example DAG that shows how users should use it.
Consider using Breeze environment for testing locally, it's a heavy docker but it ships with a working Airflow and a lot of integrations.
Be patient and persistent. It might take some time to get a review or get the final approval from Committers.
Please follow ASF Code of Conduct for all communication including (but not limited to) comments on Pull Requests, Mailing list and Slack.
Be sure to read the Airflow Coding style.
Always keep your Pull Requests rebased, otherwise your build might fail due to changes not related to your commits.
Apache Airflow is a community-driven project and together we are making it better 🚀.
In case of doubts contact the developers at:
Mailing List: dev@airflow.apache.org
Slack: https://s.apache.org/airflow-slack

jroachgolf84 · 2025-03-28T11:25:50Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

jroachgolf84 · 2025-03-28T20:49:15Z

There were a few other places where I think we might need to make changes.

dheerajkumar-solanki · 2025-03-29T00:39:09Z

There were a few other places where I think we might need to make changes.

Which places, can you provide some context, so I can update the PR.

dheerajkumar-solanki · 2025-03-29T00:56:59Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

Yes, I was able to run the failed DAG.

ghost · 2025-03-31T13:49:05Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").

https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

jroachgolf84 · 2025-04-04T16:26:29Z

@dheerajkumar-solanki, how are things going on your side? Have those unit-tests been run?

dheerajkumar-solanki · 2025-04-06T05:58:17Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").
https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

I checked the code and previous MR, there is no change required as we have not change the function definition of databricks_task_key.

All the databricks unit test files are running fine.

dheerajkumar-solanki · 2025-04-06T06:31:51Z

@pankajkoti @jroachgolf84 One of the tests pipeline is failing Tests / CI image checks / MyPy checks (mypy-providers) (pull_request), doesn't look like from the changes in this PR. Is it just due to my branch is not up to date with base branch or any other issue?

Also, there is another MR on this same issue (#48623), I like his solution but I think some pieces are missing in the MR. what would be next step in this MR now?

potiuk · 2025-04-09T00:19:24Z

rebase first and make it green

dheerajkumar-solanki · 2025-04-09T00:51:48Z

rebase first and make it green

@potiuk rebase is done, waiting for workflow approval to run the test again.

boring-cyborg · 2025-04-13T12:21:34Z

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

potiuk · 2025-04-13T12:21:39Z

Thanks!

The runtime fix for issue apache#47614 shipped in PR apache#48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: apache#47614

…66681) * Lock in DatabricksTaskBaseOperator depends_on parent-key behavior The runtime fix for issue #47614 shipped in PR #48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: #47614 * Remove manually added databricks changelog entry Provider changelogs are regenerated from git log by the release manager and should not be edited by hand. * Assert depends_on reaches the Databricks Jobs API payload Existing TestWorkflowDependsOn coverage verified the in-process create_workflow_json() output. The new TestWorkflowDependsOnWirePayload class drives _CreateDatabricksWorkflowOperator._create_or_reset_job end to end and asserts on the spec that DatabricksHook.create_job / DatabricksHook.reset_job actually receive, i.e. the payload that lands in /api/2.1/jobs/create and /api/2.1/jobs/reset. Both branches (no existing job -> create_job, existing job -> reset_job) are exercised; both assert tasks[child].depends_on == [{task_key: md5(parent)}] and tasks[parent].depends_on == []. * Fix Databricks provider mypy timezone imports * Fix Databricks provider timezone import compatibility

dheerajkumar-solanki · 2025-03-28T11:11:03Z

Fixing the DatabricksNotebookOperator invalid dependency graph issue, this comes when dependencies are defined in DatabricksWorkflowTaskGroup. Issue is child task DatabricksNotebookOperator and DatabricksTaskOperator task referencing itself. Issue in _generate_databricks_task_key function which is incorrectly generating depends_on task value.

closes: #47983

^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in airflow-core/newsfragments.

boring-cyborg · 2025-03-28T11:11:08Z

Congratulations on your first Pull Request and welcome to the Apache Airflow community! If you have any issues or are unsure about any anything please check our Contributors' Guide (https://github.com/apache/airflow/blob/main/contributing-docs/README.rst)
Here are some useful points:

Pay attention to the quality of your code (ruff, mypy and type annotations). Our pre-commits will help you with that.
In case of a new feature add useful documentation (in docstrings or in docs/ directory). Adding a new operator? Check this short guide Consider adding an example DAG that shows how users should use it.
Consider using Breeze environment for testing locally, it's a heavy docker but it ships with a working Airflow and a lot of integrations.
Be patient and persistent. It might take some time to get a review or get the final approval from Committers.
Please follow ASF Code of Conduct for all communication including (but not limited to) comments on Pull Requests, Mailing list and Slack.
Be sure to read the Airflow Coding style.
Always keep your Pull Requests rebased, otherwise your build might fail due to changes not related to your commits.
Apache Airflow is a community-driven project and together we are making it better 🚀.
In case of doubts contact the developers at:
Mailing List: dev@airflow.apache.org
Slack: https://s.apache.org/airflow-slack

jroachgolf84 · 2025-03-28T11:25:50Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

jroachgolf84 · 2025-03-28T20:49:15Z

There were a few other places where I think we might need to make changes.

dheerajkumar-solanki · 2025-03-29T00:39:09Z

There were a few other places where I think we might need to make changes.

Which places, can you provide some context, so I can update the PR.

dheerajkumar-solanki · 2025-03-29T00:56:59Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

Yes, I was able to run the failed DAG.

ghost · 2025-03-31T13:49:05Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").

https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

jroachgolf84 · 2025-04-04T16:26:29Z

@dheerajkumar-solanki, how are things going on your side? Have those unit-tests been run?

dheerajkumar-solanki · 2025-04-06T05:58:17Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").
https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

I checked the code and previous MR, there is no change required as we have not change the function definition of databricks_task_key.

All the databricks unit test files are running fine.

dheerajkumar-solanki · 2025-04-06T06:31:51Z

@pankajkoti @jroachgolf84 One of the tests pipeline is failing Tests / CI image checks / MyPy checks (mypy-providers) (pull_request), doesn't look like from the changes in this PR. Is it just due to my branch is not up to date with base branch or any other issue?

Also, there is another MR on this same issue (#48623), I like his solution but I think some pieces are missing in the MR. what would be next step in this MR now?

potiuk · 2025-04-09T00:19:24Z

rebase first and make it green

dheerajkumar-solanki · 2025-04-09T00:51:48Z

rebase first and make it green

@potiuk rebase is done, waiting for workflow approval to run the test again.

boring-cyborg · 2025-04-13T12:21:34Z

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

potiuk · 2025-04-13T12:21:39Z

Thanks!

The runtime fix for issue apache#47614 shipped in PR apache#48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: apache#47614

…66681) * Lock in DatabricksTaskBaseOperator depends_on parent-key behavior The runtime fix for issue #47614 shipped in PR #48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: #47614 * Remove manually added databricks changelog entry Provider changelogs are regenerated from git log by the release manager and should not be edited by hand. * Assert depends_on reaches the Databricks Jobs API payload Existing TestWorkflowDependsOn coverage verified the in-process create_workflow_json() output. The new TestWorkflowDependsOnWirePayload class drives _CreateDatabricksWorkflowOperator._create_or_reset_job end to end and asserts on the spec that DatabricksHook.create_job / DatabricksHook.reset_job actually receive, i.e. the payload that lands in /api/2.1/jobs/create and /api/2.1/jobs/reset. Both branches (no existing job -> create_job, existing job -> reset_job) are exercised; both assert tasks[child].depends_on == [{task_key: md5(parent)}] and tasks[parent].depends_on == []. * Fix Databricks provider mypy timezone imports * Fix Databricks provider timezone import compatibility

dheerajkumar-solanki · 2025-03-28T11:11:03Z

Fixing the DatabricksNotebookOperator invalid dependency graph issue, this comes when dependencies are defined in DatabricksWorkflowTaskGroup. Issue is child task DatabricksNotebookOperator and DatabricksTaskOperator task referencing itself. Issue in _generate_databricks_task_key function which is incorrectly generating depends_on task value.

closes: #47983

^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in airflow-core/newsfragments.

boring-cyborg · 2025-03-28T11:11:08Z

Congratulations on your first Pull Request and welcome to the Apache Airflow community! If you have any issues or are unsure about any anything please check our Contributors' Guide (https://github.com/apache/airflow/blob/main/contributing-docs/README.rst)
Here are some useful points:

Pay attention to the quality of your code (ruff, mypy and type annotations). Our pre-commits will help you with that.
In case of a new feature add useful documentation (in docstrings or in docs/ directory). Adding a new operator? Check this short guide Consider adding an example DAG that shows how users should use it.
Consider using Breeze environment for testing locally, it's a heavy docker but it ships with a working Airflow and a lot of integrations.
Be patient and persistent. It might take some time to get a review or get the final approval from Committers.
Please follow ASF Code of Conduct for all communication including (but not limited to) comments on Pull Requests, Mailing list and Slack.
Be sure to read the Airflow Coding style.
Always keep your Pull Requests rebased, otherwise your build might fail due to changes not related to your commits.
Apache Airflow is a community-driven project and together we are making it better 🚀.
In case of doubts contact the developers at:
Mailing List: dev@airflow.apache.org
Slack: https://s.apache.org/airflow-slack

jroachgolf84 · 2025-03-28T11:25:50Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

jroachgolf84 · 2025-03-28T20:49:15Z

There were a few other places where I think we might need to make changes.

dheerajkumar-solanki · 2025-03-29T00:39:09Z

There were a few other places where I think we might need to make changes.

Which places, can you provide some context, so I can update the PR.

dheerajkumar-solanki · 2025-03-29T00:56:59Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

Yes, I was able to run the failed DAG.

ghost · 2025-03-31T13:49:05Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").

https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

jroachgolf84 · 2025-04-04T16:26:29Z

@dheerajkumar-solanki, how are things going on your side? Have those unit-tests been run?

dheerajkumar-solanki · 2025-04-06T05:58:17Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").
https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

I checked the code and previous MR, there is no change required as we have not change the function definition of databricks_task_key.

All the databricks unit test files are running fine.

dheerajkumar-solanki · 2025-04-06T06:31:51Z

@pankajkoti @jroachgolf84 One of the tests pipeline is failing Tests / CI image checks / MyPy checks (mypy-providers) (pull_request), doesn't look like from the changes in this PR. Is it just due to my branch is not up to date with base branch or any other issue?

Also, there is another MR on this same issue (#48623), I like his solution but I think some pieces are missing in the MR. what would be next step in this MR now?

potiuk · 2025-04-09T00:19:24Z

rebase first and make it green

dheerajkumar-solanki · 2025-04-09T00:51:48Z

rebase first and make it green

@potiuk rebase is done, waiting for workflow approval to run the test again.

boring-cyborg · 2025-04-13T12:21:34Z

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

potiuk · 2025-04-13T12:21:39Z

Thanks!

The runtime fix for issue apache#47614 shipped in PR apache#48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: apache#47614

…66681) * Lock in DatabricksTaskBaseOperator depends_on parent-key behavior The runtime fix for issue #47614 shipped in PR #48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: #47614 * Remove manually added databricks changelog entry Provider changelogs are regenerated from git log by the release manager and should not be edited by hand. * Assert depends_on reaches the Databricks Jobs API payload Existing TestWorkflowDependsOn coverage verified the in-process create_workflow_json() output. The new TestWorkflowDependsOnWirePayload class drives _CreateDatabricksWorkflowOperator._create_or_reset_job end to end and asserts on the spec that DatabricksHook.create_job / DatabricksHook.reset_job actually receive, i.e. the payload that lands in /api/2.1/jobs/create and /api/2.1/jobs/reset. Both branches (no existing job -> create_job, existing job -> reset_job) are exercised; both assert tasks[child].depends_on == [{task_key: md5(parent)}] and tasks[parent].depends_on == []. * Fix Databricks provider mypy timezone imports * Fix Databricks provider timezone import compatibility

dheerajkumar-solanki · 2025-03-28T11:11:03Z

Fixing the DatabricksNotebookOperator invalid dependency graph issue, this comes when dependencies are defined in DatabricksWorkflowTaskGroup. Issue is child task DatabricksNotebookOperator and DatabricksTaskOperator task referencing itself. Issue in _generate_databricks_task_key function which is incorrectly generating depends_on task value.

closes: #47983

^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in airflow-core/newsfragments.

boring-cyborg · 2025-03-28T11:11:08Z

Congratulations on your first Pull Request and welcome to the Apache Airflow community! If you have any issues or are unsure about any anything please check our Contributors' Guide (https://github.com/apache/airflow/blob/main/contributing-docs/README.rst)
Here are some useful points:

Pay attention to the quality of your code (ruff, mypy and type annotations). Our pre-commits will help you with that.
In case of a new feature add useful documentation (in docstrings or in docs/ directory). Adding a new operator? Check this short guide Consider adding an example DAG that shows how users should use it.
Consider using Breeze environment for testing locally, it's a heavy docker but it ships with a working Airflow and a lot of integrations.
Be patient and persistent. It might take some time to get a review or get the final approval from Committers.
Please follow ASF Code of Conduct for all communication including (but not limited to) comments on Pull Requests, Mailing list and Slack.
Be sure to read the Airflow Coding style.
Always keep your Pull Requests rebased, otherwise your build might fail due to changes not related to your commits.
Apache Airflow is a community-driven project and together we are making it better 🚀.
In case of doubts contact the developers at:
Mailing List: dev@airflow.apache.org
Slack: https://s.apache.org/airflow-slack

jroachgolf84 · 2025-03-28T11:25:50Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

jroachgolf84 · 2025-03-28T20:49:15Z

There were a few other places where I think we might need to make changes.

dheerajkumar-solanki · 2025-03-29T00:39:09Z

There were a few other places where I think we might need to make changes.

Which places, can you provide some context, so I can update the PR.

dheerajkumar-solanki · 2025-03-29T00:56:59Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

Yes, I was able to run the failed DAG.

ghost · 2025-03-31T13:49:05Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").

https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

jroachgolf84 · 2025-04-04T16:26:29Z

@dheerajkumar-solanki, how are things going on your side? Have those unit-tests been run?

dheerajkumar-solanki · 2025-04-06T05:58:17Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").
https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

I checked the code and previous MR, there is no change required as we have not change the function definition of databricks_task_key.

All the databricks unit test files are running fine.

dheerajkumar-solanki · 2025-04-06T06:31:51Z

@pankajkoti @jroachgolf84 One of the tests pipeline is failing Tests / CI image checks / MyPy checks (mypy-providers) (pull_request), doesn't look like from the changes in this PR. Is it just due to my branch is not up to date with base branch or any other issue?

Also, there is another MR on this same issue (#48623), I like his solution but I think some pieces are missing in the MR. what would be next step in this MR now?

potiuk · 2025-04-09T00:19:24Z

rebase first and make it green

dheerajkumar-solanki · 2025-04-09T00:51:48Z

rebase first and make it green

@potiuk rebase is done, waiting for workflow approval to run the test again.

boring-cyborg · 2025-04-13T12:21:34Z

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

potiuk · 2025-04-13T12:21:39Z

Thanks!

The runtime fix for issue apache#47614 shipped in PR apache#48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: apache#47614

…66681) * Lock in DatabricksTaskBaseOperator depends_on parent-key behavior The runtime fix for issue #47614 shipped in PR #48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: #47614 * Remove manually added databricks changelog entry Provider changelogs are regenerated from git log by the release manager and should not be edited by hand. * Assert depends_on reaches the Databricks Jobs API payload Existing TestWorkflowDependsOn coverage verified the in-process create_workflow_json() output. The new TestWorkflowDependsOnWirePayload class drives _CreateDatabricksWorkflowOperator._create_or_reset_job end to end and asserts on the spec that DatabricksHook.create_job / DatabricksHook.reset_job actually receive, i.e. the payload that lands in /api/2.1/jobs/create and /api/2.1/jobs/reset. Both branches (no existing job -> create_job, existing job -> reset_job) are exercised; both assert tasks[child].depends_on == [{task_key: md5(parent)}] and tasks[parent].depends_on == []. * Fix Databricks provider mypy timezone imports * Fix Databricks provider timezone import compatibility

dheerajkumar-solanki · 2025-03-28T11:11:03Z

Fixing the DatabricksNotebookOperator invalid dependency graph issue, this comes when dependencies are defined in DatabricksWorkflowTaskGroup. Issue is child task DatabricksNotebookOperator and DatabricksTaskOperator task referencing itself. Issue in _generate_databricks_task_key function which is incorrectly generating depends_on task value.

closes: #47983

^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in airflow-core/newsfragments.

boring-cyborg · 2025-03-28T11:11:08Z

Congratulations on your first Pull Request and welcome to the Apache Airflow community! If you have any issues or are unsure about any anything please check our Contributors' Guide (https://github.com/apache/airflow/blob/main/contributing-docs/README.rst)
Here are some useful points:

Pay attention to the quality of your code (ruff, mypy and type annotations). Our pre-commits will help you with that.
In case of a new feature add useful documentation (in docstrings or in docs/ directory). Adding a new operator? Check this short guide Consider adding an example DAG that shows how users should use it.
Consider using Breeze environment for testing locally, it's a heavy docker but it ships with a working Airflow and a lot of integrations.
Be patient and persistent. It might take some time to get a review or get the final approval from Committers.
Please follow ASF Code of Conduct for all communication including (but not limited to) comments on Pull Requests, Mailing list and Slack.
Be sure to read the Airflow Coding style.
Always keep your Pull Requests rebased, otherwise your build might fail due to changes not related to your commits.
Apache Airflow is a community-driven project and together we are making it better 🚀.
In case of doubts contact the developers at:
Mailing List: dev@airflow.apache.org
Slack: https://s.apache.org/airflow-slack

jroachgolf84 · 2025-03-28T11:25:50Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

jroachgolf84 · 2025-03-28T20:49:15Z

There were a few other places where I think we might need to make changes.

dheerajkumar-solanki · 2025-03-29T00:39:09Z

There were a few other places where I think we might need to make changes.

Which places, can you provide some context, so I can update the PR.

dheerajkumar-solanki · 2025-03-29T00:56:59Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

Yes, I was able to run the failed DAG.

ghost · 2025-03-31T13:49:05Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").

https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

jroachgolf84 · 2025-04-04T16:26:29Z

@dheerajkumar-solanki, how are things going on your side? Have those unit-tests been run?

dheerajkumar-solanki · 2025-04-06T05:58:17Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").
https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

I checked the code and previous MR, there is no change required as we have not change the function definition of databricks_task_key.

All the databricks unit test files are running fine.

dheerajkumar-solanki · 2025-04-06T06:31:51Z

@pankajkoti @jroachgolf84 One of the tests pipeline is failing Tests / CI image checks / MyPy checks (mypy-providers) (pull_request), doesn't look like from the changes in this PR. Is it just due to my branch is not up to date with base branch or any other issue?

Also, there is another MR on this same issue (#48623), I like his solution but I think some pieces are missing in the MR. what would be next step in this MR now?

potiuk · 2025-04-09T00:19:24Z

rebase first and make it green

dheerajkumar-solanki · 2025-04-09T00:51:48Z

rebase first and make it green

@potiuk rebase is done, waiting for workflow approval to run the test again.

boring-cyborg · 2025-04-13T12:21:34Z

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

potiuk · 2025-04-13T12:21:39Z

Thanks!

The runtime fix for issue apache#47614 shipped in PR apache#48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: apache#47614

…66681) * Lock in DatabricksTaskBaseOperator depends_on parent-key behavior The runtime fix for issue #47614 shipped in PR #48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: #47614 * Remove manually added databricks changelog entry Provider changelogs are regenerated from git log by the release manager and should not be edited by hand. * Assert depends_on reaches the Databricks Jobs API payload Existing TestWorkflowDependsOn coverage verified the in-process create_workflow_json() output. The new TestWorkflowDependsOnWirePayload class drives _CreateDatabricksWorkflowOperator._create_or_reset_job end to end and asserts on the spec that DatabricksHook.create_job / DatabricksHook.reset_job actually receive, i.e. the payload that lands in /api/2.1/jobs/create and /api/2.1/jobs/reset. Both branches (no existing job -> create_job, existing job -> reset_job) are exercised; both assert tasks[child].depends_on == [{task_key: md5(parent)}] and tasks[parent].depends_on == []. * Fix Databricks provider mypy timezone imports * Fix Databricks provider timezone import compatibility

dheerajkumar-solanki · 2025-03-28T11:11:03Z

Fixing the DatabricksNotebookOperator invalid dependency graph issue, this comes when dependencies are defined in DatabricksWorkflowTaskGroup. Issue is child task DatabricksNotebookOperator and DatabricksTaskOperator task referencing itself. Issue in _generate_databricks_task_key function which is incorrectly generating depends_on task value.

closes: #47983

^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in airflow-core/newsfragments.

boring-cyborg · 2025-03-28T11:11:08Z

Congratulations on your first Pull Request and welcome to the Apache Airflow community! If you have any issues or are unsure about any anything please check our Contributors' Guide (https://github.com/apache/airflow/blob/main/contributing-docs/README.rst)
Here are some useful points:

Pay attention to the quality of your code (ruff, mypy and type annotations). Our pre-commits will help you with that.
In case of a new feature add useful documentation (in docstrings or in docs/ directory). Adding a new operator? Check this short guide Consider adding an example DAG that shows how users should use it.
Consider using Breeze environment for testing locally, it's a heavy docker but it ships with a working Airflow and a lot of integrations.
Be patient and persistent. It might take some time to get a review or get the final approval from Committers.
Please follow ASF Code of Conduct for all communication including (but not limited to) comments on Pull Requests, Mailing list and Slack.
Be sure to read the Airflow Coding style.
Always keep your Pull Requests rebased, otherwise your build might fail due to changes not related to your commits.
Apache Airflow is a community-driven project and together we are making it better 🚀.
In case of doubts contact the developers at:
Mailing List: dev@airflow.apache.org
Slack: https://s.apache.org/airflow-slack

jroachgolf84 · 2025-03-28T11:25:50Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

jroachgolf84 · 2025-03-28T20:49:15Z

There were a few other places where I think we might need to make changes.

dheerajkumar-solanki · 2025-03-29T00:39:09Z

There were a few other places where I think we might need to make changes.

Which places, can you provide some context, so I can update the PR.

dheerajkumar-solanki · 2025-03-29T00:56:59Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

Yes, I was able to run the failed DAG.

ghost · 2025-03-31T13:49:05Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").

https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

jroachgolf84 · 2025-04-04T16:26:29Z

@dheerajkumar-solanki, how are things going on your side? Have those unit-tests been run?

dheerajkumar-solanki · 2025-04-06T05:58:17Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").
https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

I checked the code and previous MR, there is no change required as we have not change the function definition of databricks_task_key.

All the databricks unit test files are running fine.

dheerajkumar-solanki · 2025-04-06T06:31:51Z

@pankajkoti @jroachgolf84 One of the tests pipeline is failing Tests / CI image checks / MyPy checks (mypy-providers) (pull_request), doesn't look like from the changes in this PR. Is it just due to my branch is not up to date with base branch or any other issue?

Also, there is another MR on this same issue (#48623), I like his solution but I think some pieces are missing in the MR. what would be next step in this MR now?

potiuk · 2025-04-09T00:19:24Z

rebase first and make it green

dheerajkumar-solanki · 2025-04-09T00:51:48Z

rebase first and make it green

@potiuk rebase is done, waiting for workflow approval to run the test again.

boring-cyborg · 2025-04-13T12:21:34Z

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

potiuk · 2025-04-13T12:21:39Z

Thanks!

The runtime fix for issue apache#47614 shipped in PR apache#48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: apache#47614

…66681) * Lock in DatabricksTaskBaseOperator depends_on parent-key behavior The runtime fix for issue #47614 shipped in PR #48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: #47614 * Remove manually added databricks changelog entry Provider changelogs are regenerated from git log by the release manager and should not be edited by hand. * Assert depends_on reaches the Databricks Jobs API payload Existing TestWorkflowDependsOn coverage verified the in-process create_workflow_json() output. The new TestWorkflowDependsOnWirePayload class drives _CreateDatabricksWorkflowOperator._create_or_reset_job end to end and asserts on the spec that DatabricksHook.create_job / DatabricksHook.reset_job actually receive, i.e. the payload that lands in /api/2.1/jobs/create and /api/2.1/jobs/reset. Both branches (no existing job -> create_job, existing job -> reset_job) are exercised; both assert tasks[child].depends_on == [{task_key: md5(parent)}] and tasks[parent].depends_on == []. * Fix Databricks provider mypy timezone imports * Fix Databricks provider timezone import compatibility

dheerajkumar-solanki · 2025-03-28T11:11:03Z

Fixing the DatabricksNotebookOperator invalid dependency graph issue, this comes when dependencies are defined in DatabricksWorkflowTaskGroup. Issue is child task DatabricksNotebookOperator and DatabricksTaskOperator task referencing itself. Issue in _generate_databricks_task_key function which is incorrectly generating depends_on task value.

closes: #47983

^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in airflow-core/newsfragments.

boring-cyborg · 2025-03-28T11:11:08Z

Congratulations on your first Pull Request and welcome to the Apache Airflow community! If you have any issues or are unsure about any anything please check our Contributors' Guide (https://github.com/apache/airflow/blob/main/contributing-docs/README.rst)
Here are some useful points:

Pay attention to the quality of your code (ruff, mypy and type annotations). Our pre-commits will help you with that.
In case of a new feature add useful documentation (in docstrings or in docs/ directory). Adding a new operator? Check this short guide Consider adding an example DAG that shows how users should use it.
Consider using Breeze environment for testing locally, it's a heavy docker but it ships with a working Airflow and a lot of integrations.
Be patient and persistent. It might take some time to get a review or get the final approval from Committers.
Please follow ASF Code of Conduct for all communication including (but not limited to) comments on Pull Requests, Mailing list and Slack.
Be sure to read the Airflow Coding style.
Always keep your Pull Requests rebased, otherwise your build might fail due to changes not related to your commits.
Apache Airflow is a community-driven project and together we are making it better 🚀.
In case of doubts contact the developers at:
Mailing List: dev@airflow.apache.org
Slack: https://s.apache.org/airflow-slack

jroachgolf84 · 2025-03-28T11:25:50Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

jroachgolf84 · 2025-03-28T20:49:15Z

There were a few other places where I think we might need to make changes.

dheerajkumar-solanki · 2025-03-29T00:39:09Z

There were a few other places where I think we might need to make changes.

Which places, can you provide some context, so I can update the PR.

dheerajkumar-solanki · 2025-03-29T00:56:59Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

Yes, I was able to run the failed DAG.

ghost · 2025-03-31T13:49:05Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").

https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

jroachgolf84 · 2025-04-04T16:26:29Z

@dheerajkumar-solanki, how are things going on your side? Have those unit-tests been run?

dheerajkumar-solanki · 2025-04-06T05:58:17Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").
https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

I checked the code and previous MR, there is no change required as we have not change the function definition of databricks_task_key.

All the databricks unit test files are running fine.

dheerajkumar-solanki · 2025-04-06T06:31:51Z

@pankajkoti @jroachgolf84 One of the tests pipeline is failing Tests / CI image checks / MyPy checks (mypy-providers) (pull_request), doesn't look like from the changes in this PR. Is it just due to my branch is not up to date with base branch or any other issue?

Also, there is another MR on this same issue (#48623), I like his solution but I think some pieces are missing in the MR. what would be next step in this MR now?

potiuk · 2025-04-09T00:19:24Z

rebase first and make it green

dheerajkumar-solanki · 2025-04-09T00:51:48Z

rebase first and make it green

@potiuk rebase is done, waiting for workflow approval to run the test again.

boring-cyborg · 2025-04-13T12:21:34Z

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

potiuk · 2025-04-13T12:21:39Z

Thanks!

The runtime fix for issue apache#47614 shipped in PR apache#48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: apache#47614

…66681) * Lock in DatabricksTaskBaseOperator depends_on parent-key behavior The runtime fix for issue #47614 shipped in PR #48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: #47614 * Remove manually added databricks changelog entry Provider changelogs are regenerated from git log by the release manager and should not be edited by hand. * Assert depends_on reaches the Databricks Jobs API payload Existing TestWorkflowDependsOn coverage verified the in-process create_workflow_json() output. The new TestWorkflowDependsOnWirePayload class drives _CreateDatabricksWorkflowOperator._create_or_reset_job end to end and asserts on the spec that DatabricksHook.create_job / DatabricksHook.reset_job actually receive, i.e. the payload that lands in /api/2.1/jobs/create and /api/2.1/jobs/reset. Both branches (no existing job -> create_job, existing job -> reset_job) are exercised; both assert tasks[child].depends_on == [{task_key: md5(parent)}] and tasks[parent].depends_on == []. * Fix Databricks provider mypy timezone imports * Fix Databricks provider timezone import compatibility

dheerajkumar-solanki · 2025-03-28T11:11:03Z

Fixing the DatabricksNotebookOperator invalid dependency graph issue, this comes when dependencies are defined in DatabricksWorkflowTaskGroup. Issue is child task DatabricksNotebookOperator and DatabricksTaskOperator task referencing itself. Issue in _generate_databricks_task_key function which is incorrectly generating depends_on task value.

closes: #47983

^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in airflow-core/newsfragments.

boring-cyborg · 2025-03-28T11:11:08Z

Congratulations on your first Pull Request and welcome to the Apache Airflow community! If you have any issues or are unsure about any anything please check our Contributors' Guide (https://github.com/apache/airflow/blob/main/contributing-docs/README.rst)
Here are some useful points:

Pay attention to the quality of your code (ruff, mypy and type annotations). Our pre-commits will help you with that.
In case of a new feature add useful documentation (in docstrings or in docs/ directory). Adding a new operator? Check this short guide Consider adding an example DAG that shows how users should use it.
Consider using Breeze environment for testing locally, it's a heavy docker but it ships with a working Airflow and a lot of integrations.
Be patient and persistent. It might take some time to get a review or get the final approval from Committers.
Please follow ASF Code of Conduct for all communication including (but not limited to) comments on Pull Requests, Mailing list and Slack.
Be sure to read the Airflow Coding style.
Always keep your Pull Requests rebased, otherwise your build might fail due to changes not related to your commits.
Apache Airflow is a community-driven project and together we are making it better 🚀.
In case of doubts contact the developers at:
Mailing List: dev@airflow.apache.org
Slack: https://s.apache.org/airflow-slack

jroachgolf84 · 2025-03-28T11:25:50Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

jroachgolf84 · 2025-03-28T20:49:15Z

There were a few other places where I think we might need to make changes.

dheerajkumar-solanki · 2025-03-29T00:39:09Z

There were a few other places where I think we might need to make changes.

Which places, can you provide some context, so I can update the PR.

dheerajkumar-solanki · 2025-03-29T00:56:59Z

@dheerajkumar-solanki, nice stuff! I was going to approach this in a bit of a different way, but I like what you did. Question; were you able to run that DAG that you originally had posted?

Yes, I was able to run the failed DAG.

ghost · 2025-03-31T13:49:05Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").

https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

jroachgolf84 · 2025-04-04T16:26:29Z

@dheerajkumar-solanki, how are things going on your side? Have those unit-tests been run?

dheerajkumar-solanki · 2025-04-06T05:58:17Z

This was one of the places I believe would need a change (and was part of the original PR that we believe may have been "breaking").
https://github.com/apache/airflow/blob/3dbed17b0ef52b3a46483219feb4b311251990eb/providers/databricks/src/airflow/providers/databricks/plugins/databricks_workflow.py#L321C1-L337C115

I checked the code and previous MR, there is no change required as we have not change the function definition of databricks_task_key.

All the databricks unit test files are running fine.

dheerajkumar-solanki · 2025-04-06T06:31:51Z

@pankajkoti @jroachgolf84 One of the tests pipeline is failing Tests / CI image checks / MyPy checks (mypy-providers) (pull_request), doesn't look like from the changes in this PR. Is it just due to my branch is not up to date with base branch or any other issue?

Also, there is another MR on this same issue (#48623), I like his solution but I think some pieces are missing in the MR. what would be next step in this MR now?

potiuk · 2025-04-09T00:19:24Z

rebase first and make it green

dheerajkumar-solanki · 2025-04-09T00:51:48Z

rebase first and make it green

@potiuk rebase is done, waiting for workflow approval to run the test again.

boring-cyborg · 2025-04-13T12:21:34Z

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

potiuk · 2025-04-13T12:21:39Z

Thanks!

The runtime fix for issue apache#47614 shipped in PR apache#48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: apache#47614

…66681) * Lock in DatabricksTaskBaseOperator depends_on parent-key behavior The runtime fix for issue #47614 shipped in PR #48492; this PR adds end-to-end regression coverage so the bug cannot silently regress, plus small type-hint and constructor-clarity follow-ups in the same area. Tests build a real DAG + DatabricksWorkflowTaskGroup with DatabricksNotebookOperator tasks and assert depends_on payloads for the default-key, custom-key, >100-char-key, diamond, fan-out, root-task, and external-upstream cases. Also fixes the existing test_convert_to_databricks_workflow_task to pass strings (not mocks) so the depends_on branch is actually exercised, and adds a one-line check that _generate_databricks_task_key raises when called with a parent task_id but no task_dict. closes: #47614 * Remove manually added databricks changelog entry Provider changelogs are regenerated from git log by the release manager and should not be edited by hand. * Assert depends_on reaches the Databricks Jobs API payload Existing TestWorkflowDependsOn coverage verified the in-process create_workflow_json() output. The new TestWorkflowDependsOnWirePayload class drives _CreateDatabricksWorkflowOperator._create_or_reset_job end to end and asserts on the spec that DatabricksHook.create_job / DatabricksHook.reset_job actually receive, i.e. the payload that lands in /api/2.1/jobs/create and /api/2.1/jobs/reset. Both branches (no existing job -> create_job, existing job -> reset_job) are exercised; both assert tasks[child].depends_on == [{task_key: md5(parent)}] and tasks[parent].depends_on == []. * Fix Databricks provider mypy timezone imports * Fix Databricks provider timezone import compatibility

boring-cyborgBot added area:providers provider:databricks labels Mar 28, 2025

dheerajkumar-solanki force-pushed the issue-47983 branch from 1de6d22 to 3b93d41Compare March 31, 2025 20:00

dheerajkumar-solanki force-pushed the issue-47983 branch from 3b93d41 to 083240eCompare April 9, 2025 00:49

dheerajkumar-solanki force-pushed the issue-47983 branch from dc0956e to 8f3f8c2Compare April 12, 2025 11:50

Fixing DatabricksNotebookOperator invalid dependency graph issue
f431994

dheerajkumar-solanki force-pushed the issue-47983 branch from 8f3f8c2 to f431994Compare April 13, 2025 02:49

potiuk approved these changes Apr 13, 2025
View reviewed changes

potiuk merged commit 9dcce2f into apache:mainApr 13, 2025

eladkal mentioned this pull request Apr 16, 2025
Status of testing Providers that were prepared on April 16, 2025 #49354
Closed
56 tasks

Vamsi-klu mentioned this pull request May 11, 2026
Lock in Databricks workflow depends_on parent-key behavior (#47614) #66681
Merged
1 task

potiuk mentioned this pull request Jun 3, 2026
Status of testing Providers that were prepared on June 02, 2026 #67957
Closed