[X33834]

Problem

Closes: #73091

BigQueryHook.create_table()'s table_resource parameter is typed and documented to accept a dict, Table, TableReference, or TableListItem, but the implementation only worked with a plain dict:

• Table instance: Table.from_api_repr() was called backwards (it expects an API-JSON dict, not a Table object) → TypeError: argument of type 'Table' is not iterable
• TableReference / TableListItem: raw **-unpacking on a non-mapping → TypeError: 'TableReference' object is not a mapping

This is reachable from BigQueryCreateTableOperator, which has the identical type hint and forwards table_resource straight into the hook. A DAG author who follows the documented type hint and passes a Table obtained from get_table() / list_tables() hits a crash.

Solution

Convert any object input to a plain dict via .to_api_repr() before merging schema_fields and building the final Table. A bare TableReference is wrapped under "tableReference" since its to_api_repr() is flat, unlike Table / TableListItem.

Tests

Added unit tests in TestTableOperations covering all four input forms (dict, Table, TableReference, TableListItem) plus schema merging:

• test_create_table_with_table_object
• test_create_table_with_table_reference_object
• test_create_table_with_table_list_item
• test_create_table_with_table_object_and_schema_fields

Ran locally: conversion logic verified against real google.cloud.bigquery objects; ruff check and py_compile clean. Full provider test suite runs in CI via breeze.

Was generative AI tooling used to co-author this PR?

• Yes (please specify the tool below)

Generated-by: DeepSeek via TraeCode AI assistant following the guidelines

---

[boring-cyborg]

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
Here are some useful points:

• Pay attention to the quality of your code (ruff, mypy and type annotations). Our prek-hooks 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

[shahar1]

Welcome to Apache Airflow!
Before letting maintainers to review your PR:

1. Please add AI attribution as part of the PR body according to the PR template.
2. We don't use conventional commits - please remove fix(google): from title.

[X33834]

Thanks @shahar1 for the review. Both points addressed:

1. Added the generative-AI disclosure section to the PR description per the PR template.
2. Removed the conventional-commit prefix from the title.

Ready for another look.

Concerns addressed

[X33834]

Friendly ping — both review points were addressed on Sep 17 (generative-AI disclosure added to the PR description, conventional-commit prefix removed from the title). GitHub checks are green.

Happy to adjust anything else — could someone take another look when you get a chance?

[molcay]

Hi @X33834,

Thank you for your PR! The change seems GDM.

In the google provider side, we want to also see the system tests results after a change.
Did you also run the system tests which are related to the BigQuery service?

If so, please share your results as well.

Since you change the BigQueryHook, it is good to run all BigQuery related system tests

Thank you in advance!

[X33834]

Thanks @molcay!

The PR's CI run includes the BigQuery unit test suite and is green (providers/google/tests/unit/google/cloud/hooks/test_bigquery.py, including the new cases for Table/TableReference/TableListItem inputs).

For the system tests under providers/google/tests/system/google/cloud/bigquery/: I don't have access to a GCP project with the BigQuery API enabled, so I couldn't run them myself. Happy to run them if credentials can be provided, or feel free to trigger them on your side if that's easier.

For what it's worth, the change is contained to input normalization in create_table: it accepts Table/TableReference/TableListItem objects and converts them via to_api_repr() (previously a Table object was passed through Table.from_api_repr, which raised). The dict path is unchanged, so existing system tests that pass dicts exercise the same code path as before.

[X33834]

Thanks for the pointer, @molcay. I don't have access to a GCP project with the BigQuery API enabled, so I can't run the system tests from here. If someone on the team can trigger providers/google/tests/system/google/cloud/bigquery/ on this branch, I'd be glad to look at any failures.

To make the change easy to reason about while those run:

• create_table now normalizes Table / TableReference / TableListItem inputs through to_api_repr() before calling the API; the dict path is unchanged, so existing system tests that pass dicts exercise the same code path as before.
• The BigQuery unit suite in this PR's CI is green, including the new cases covering each accepted input type.

Happy to adjust the implementation if the system tests surface anything.

[molcay]

Hi @X33834,

I am trying to run the system tests for bigquery on your branch, when I have a result I will let you know.

[X33834]

Hi @molcay, thanks a lot for running the BigQuery system tests on my branch — really appreciate it!

Quick note: the Static checks failure comes from the breeze doc-image regeneration hooks (dev/breeze/doc/images/output_workflow-run*), unrelated to the actual PR changes. I've updated the branch with the latest main to clear it.

Heads-up: the CI workflows on the new commit are currently in action_required state — would you mind re-approving/triggering them when you get a chance? Looking forward to the system test results. 🙏

[molcay]

Hi @X33834,

It seems like your change did not break anything.

Since your change is about handling different kind of table definitions for the BigQueryCreateTableOperator; we must also add more tasks to the bigquery_tables system test.
As you can (see)[https://github.com/X33834/airflow/blob/fix-bigquery-create-table/providers/google/tests/system/google/cloud/bigquery/example_bigquery_tables.py#L85-L97], we already have one create_table task with dict config for table_resource argument. We must add the other type example in this DAG.

Can you make this change as well to make the system test future-proofed?

Also, please remove the newsfragment file.

[X33834]

Both pushed to the branch:

• Added a create_table_from_table_object task to example_bigquery_tables.py passing a google.cloud.bigquery.Table instance (same schema as the dict example), wired into the chain right after create_table, so both supported input types of BigQueryHook.create_table stay covered.
• Removed the newsfragment.

CI is running; happy to add a TableListItem example too if you think it's worth covering.

[potiuk]

Approving. Passing a Table used to crash (Table.from_api_repr(Table) and then **Table), and the new path handles both Table and a bare TableReference correctly. to_api_repr() returns a copy, so the caller's object isn't modified. I pushed a fixup that only moves the google.cloud.bigquery import into the third-party import block, since ruff's import-order check flagged it.

---

Drafted-by: Claude Code (Opus 5.5); reviewed by @potiuk before posting

[boring-cyborg]

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