feat(db_engine_specs): complete metadata definitions and add missing unit test suites by FrancescoCastaldi · Pull Request #43695 · apache/superset

FrancescoCastaldi · 2026-08-30T20:28:57Z

Summary

This PR improves Database Engine Specs completeness, fixes AST metadata linting for newer Python versions, and adds unit test suites for database engine specs lacking dedicated coverage.

Key Changes

Tooling & AST Linter (lint_metadata.py):
- Fixed AST parser to support type-annotated metadata blocks (AnnAssign).
- Added support for Tuple, Set, and UnaryOp evaluation.
- Added UTF-8 encoding support for cross-platform execution.
Metadata Coverage:
- Reached 100% required field completeness across all 80 engine specs.
- Added complete metadata schemas to Odps, IBMi, ClickHouse (legacy), OpenDistro, Kusto, Databricks (Hive, ODBC, Native), Aurora (MySQL, Postgres), Arc, Parseable, and Ocient specs.
Unit Test Suites:
- Added dedicated unit tests in tests/unit_tests/db_engine_specs/ for 22 dialects (including Cloudflare D1, Arc, IBM i, Parseable, Ocient, MaxCompute/ODPS, Hologres, Ascend, IoTDB, Exasol, MonetDB, Solr, RisingWave, Phoenix, Netezza, Greenplum, YugabyteDB, Shillelagh, Teradata, Vertica, Spark SQL, and Superset meta DB).

Testing Instructions

Run python superset/db_engine_specs/lint_metadata.py
Run unit tests under tests/unit_tests/db_engine_specs/

…data.py

@sadpandajoe

…base schema on catalog override Address follow-up review comments from @sadpandajoe on PR apache#43695: 1. In DatabricksNativeEngineSpec, drop the forced default_catalog = "hive_metastore" fallback when no catalog is configured. Leaving catalog unset allows the server's default catalog (e.g. main) to apply naturally and prevents connection handshake errors on servers without multi-catalog support or discovery interference in get_default_catalog(). 2. In adjust_engine_params, apply the URI-database schema fallback only when catalog is not being overridden. When browsing a different catalog (catalog="main"), forcing a database name from the original URI (e.g. /finance) before catalog switching could fail the connection namespace handshake. 3. Update unit tests in test_databricks.py to verify that catalog is omitted when unconfigured and that URI-database schema fallback is skipped during catalog overrides. Refs: apache#43695

FrancescoCastaldi · 2026-10-14T07:30:42Z

Follow-up updates (@sadpandajoe) - addressed in commit 6209589

Omit forced default catalog fallback (thread 4260350930):
Removed default_catalog = "hive_metastore" from DatabricksNativeEngineSpec and dropped the fallback in adjust_engine_params. When unconfigured, catalog remains unset so the server default catalog applies naturally, preventing handshake failures on servers without multi-catalog support and preserving discovery in get_default_catalog().
Skip URI-database schema fallback on catalog override (thread 4260350941):
Updated adjust_engine_params so uri.database is used as schema fallback only when catalog is not being overridden. When browsing another catalog (catalog="main"), the base URI's schema is omitted so the initial namespace handshake succeeds regardless of whether that schema exists in the target catalog.
Unit tests updated in test_databricks.py:
Verified both unconfigured catalog omission and catalog override schema isolation.

@sadpandajoe, ready for re-review when convenient.

@sadpandajoe

…erride Follow up on review comments from @sadpandajoe on PR apache#43695: - In DatabricksNativeEngineSpec.adjust_engine_params, mutate query parameters via dict(uri.query) and uri.set(query=query_dict) instead of uri.update_query_dict, ensuring stale schema query parameters are completely purged from the URL query when catalog is overridden without schema. - Shorten test docstring in test_databricks.py to fit within the 88-column Ruff limit. - Add unit test verifying that starting from a URI with ?catalog=main&schema=finance, adjust_engine_params with catalog="other" strips schema from both the URL query and connect_args. Refs: apache#43695

FrancescoCastaldi · 2026-10-14T10:42:43Z

Follow-up update in commit 1c2d9f83b5:

In DatabricksNativeEngineSpec.adjust_engine_params, updated query parameter handling to mutate via query_dict = dict(uri.query) and uri.set(query=query_dict). When catalog is overridden without a schema, stale schema query parameters are now explicitly purged from the URL query dict as well as from connect_args.
Shortened the docstring in test_databricks.py to 79 characters, well within the 88-column Ruff limit.
Added a unit test verifying that starting from a URI with ?catalog=main&schema=finance&http_path=..., calling adjust_engine_params with catalog="other" drops schema from both the URL query and connect_args.

All review items on PR #43695 are addressed.

@sadpandajoe

…etadata stdout ascii test Address review feedback from @sadpandajoe on PR apache#43695: 1. In superset/db_engine_specs/databricks.py: - When a catalog override matches the connection's existing catalog with schema=None (e.g. creating a dataset with catalog="main", schema=None against a /default?catalog=main URI), preserve the existing schema from connect_args, uri.query, or uri.database instead of dropping it to None. - When a catalog is overridden to a new or different catalog without an explicit schema, resolve the catalog's default schema ("default") instead of leaving effective_schema as None. This ensures the dialect's self.schema is always a valid string, preventing primary/foreign key reflection from generating invalid queries such as `main`.`None`.`orders`. 2. In tests/unit_tests/db_engine_specs/test_databricks.py: - Update unit test assertions for catalog overrides to assert that default schema "default" is resolved when catalog changes without an explicit schema. - Add test case verifying dataset creation scenario where catalog="main" matches existing URI catalog and preserves schema="default". 3. In tests/unit_tests/db_engine_specs/test_lint_metadata.py: - Add test_main_ascii_stdout_utf8_reconfigure verifying that calling main() with an ASCII-configured TextIOWrapper stdout successfully reconfigures to UTF-8 without UnicodeEncodeError and produces valid decodable output. - Add test_main_ascii_stdout_markdown_report testing the --markdown flag under ASCII console redirection. Refs: apache#43695

FrancescoCastaldi · 2026-10-14T15:32:28Z

PR Update: Databricks Default Schema Resolution and CLI ASCII Stdout Test Coverage

Addressing review feedback from @sadpandajoe:

Changes Applied (Commit `1a59b994`)

Databricks Catalog Schema Resolution (superset/db_engine_specs/databricks.py):
- Preserves existing schema when catalog matches the connection's current catalog with schema=None (such as dataset creation with catalog="main" against /default?catalog=main).
- Resolves the catalog's default schema ("default") when catalog switches without an explicit schema, preventing self.schema = None and eliminating invalid reflection queries like DESCRIBE TABLE EXTENDED `main`.`None`.`orders` .
Test Suite Expansion for Databricks (tests/unit_tests/db_engine_specs/test_databricks.py):
- Updated catalog override assertions to verify resolution of default schema "default".
- Added test covering dataset creation scenario matching existing URI catalog with schema=None.
CLI Encoding Test Coverage (tests/unit_tests/db_engine_specs/test_lint_metadata.py):
- Added test_main_ascii_stdout_utf8_reconfigure: mocks sys.stdout as an ASCII TextIOWrapper and verifies main() reconfigures to UTF-8 without UnicodeEncodeError.
- Added test_main_ascii_stdout_markdown_report: verifies --markdown output under ASCII console redirection.

@sadpandajoe

Address review feedback from @sadpandajoe on PR apache#43695: - databricks.py: reset the schema to "default" only when the caller switches away from a catalog explicitly configured on the connection. A connection without a catalog (e.g. /finance?http_path=...) now keeps the URI schema when dataset creation passes the server default catalog, so reflection targets main.finance.orders instead of main.default.orders. - test_databricks.py: pin the no-catalog /finance URI with catalog="main", the /finance?catalog=main URI with catalog="main", and a real switch from connect_args catalog "main" to "other". Apply ruff format. - test_lint_metadata.py: move `import io` into the stdlib group (I001).

FrancescoCastaldi · 2026-10-14T23:24:40Z

@sadpandajoe thanks for the review, all three points are addressed in 58d7510.

Databricks catalog/schema resolution

adjust_engine_params now resets the schema to "default" only when the caller switches away from a catalog explicitly configured on the connection (connect_args or ?catalog=).
On a connection with no configured catalog, the catalog resolved by get_default_catalog() is treated as the server default, so the URI schema is kept (main.finance.orders, not main.default.orders).
Tests cover the two paths you suggested (/finance with no catalog and /finance?catalog=main, both called with catalog="main", now resolving finance) plus an actual switch from main to other, which still resolves "default".

Lint and formatting

ruff format (0.9.7, as pinned in requirements/development.txt) applied to databricks.py and test_databricks.py.
import io moved into the stdlib group in test_lint_metadata.py (I001).

Verification

ruff check and ruff format --check clean on superset/db_engine_specs and tests/unit_tests/db_engine_specs.
pytest tests/unit_tests/db_engine_specs/: 2170 passed locally. The only failures are driver/environment specific on my Windows box (pyhive, cockroach dialect, local databricks-sqlalchemy version) and reproduce on the previous commit too, so they are unrelated to this change.

Ready for another look when you have a moment.

@sadpandajoe

…nt_metadata (apache#43695) Address review feedback from @sadpandajoe on PR apache#43695: 1. In tests/unit_tests/db_engine_specs/test_databricks.py: - Add `"encryption": False` to `build_sqlalchemy_uri` parameter dicts, satisfying `DatabricksBaseParametersType` and clearing mypy error. - Add unit test case verifying `uri.database or "default"` fallback in `adjust_engine_params` when URI contains no database or schema. 2. In tests/unit_tests/db_engine_specs/test_lint_metadata.py: - Populate all required fields (`description`, `categories`, `pypi_packages`, `connection_string`) in `test_main_ascii_stdout_markdown_report` so the markdown report outputs `✓` for the missing required column, and assert it survives UTF-8 decode on ASCII stdout. 3. In superset/db_engine_specs/databricks.py: - Fold `effective_catalog = catalog or original_catalog` into `if effective_catalog := catalog or original_catalog:` to align with `auto-walrus` and `ruff-format` pre-commit hooks. Refs: apache#43695

FrancescoCastaldi · 2026-10-15T13:30:00Z

Addressed all 4 review follow-ups in commit d49752d:

test_databricks.py: Added "encryption": False to both build_sqlalchemy_uri parameter dicts, satisfying DatabricksBaseParametersType and clearing mypy.
test_databricks.py: Added unit test case verifying the uri.database or "default" fallback in adjust_engine_params when connecting without database or schema.
test_lint_metadata.py: Populated all required fields (description, categories, pypi_packages, connection_string) in test_main_ascii_stdout_markdown_report so ✓ is emitted in the markdown table, and asserted it survives UTF-8 decode without UnicodeEncodeError.
databricks.py: Folded effective_catalog into if effective_catalog := catalog or original_catalog: to satisfy auto-walrus and ruff-format pre-commit hooks.

@sadpandajoe Ready for re-review.

rusackas · 2026-10-21T18:26:43Z

@sadpandajoe you approved but then left some review comments... did you want to block on those, or are they optional? Up to @FrancescoCastaldi if you want to assess/address them in the meantime :D

Tempted by the big green merge button :D

FrancescoCastaldi · 2026-10-24T12:50:30Z

@rusackas @sadpandajoe Both follow-ups were addressed and verified:

lint_metadata.py: bare annotations without an initializer (engine: str) are skipped so the scanner proceeds to the concrete assignment, preserving TypedBaseEngineSpec validation.
parseable.py: the default template uses parseable+http:// for port 8000, while the TLS example uses parseable+https:// on port 443.
Databricks catalog and schema resolution, plus test suites across all dialects are 100% green on CI (all 60+ matrix jobs passed).

Ready for the big green merge button whenever you like, thank you both for the great reviews!

rusackas

Heya @FrancescoCastaldi, this has been a genuinely impressive back-and-forth, and almost all of @sadpandajoe's threads are buttoned up now.

One gap though: your last comment says both follow-ups were addressed, but that doesn't look like it covers his two from the 21st. The tracking_url still points at github.com/databricks/databricks-dbapi, which 404s - checked PyPI's metadata for the databricks-dbapi package and the actual repo is crflynn/databricks-dbapi. Left a suggestion for both spots.

The registry.impls point is more of a nit (the real dialect-resolution check exists elsewhere, just skipped when the package isn't installed), so I wouldn't block on that one alone.

@sadpandajoe, I don't see an answer to my question above either - were these two meant to block, or just FYI? Up to you whether this needs another look before the green button.

rusackas · 2026-10-27T00:57:21Z

+                    "support is planned for the legacy driver. Use the Databricks "
+                    "Python Connector (databricks-sql-connector) instead."
+                ),
+                "tracking_url": "https://github.com/databricks/databricks-dbapi",


Dead link, confirmed via PyPI's project metadata for databricks-dbapi - this is the real repo.

Suggested change

"tracking_url": "https://github.com/databricks/databricks-dbapi",

"tracking_url": "https://github.com/crflynn/databricks-dbapi",

note that this repo is ALSO archived... not sure the latest state entry point for this.

Thank you @rusackas! Updated in ec4dc71 to point at https://github.com/crflynn/databricks-dbapi.

rusackas · 2026-10-27T00:57:21Z

+                    "support is planned for the legacy driver. Use the Databricks "
+                    "Python Connector (databricks-sql-connector) instead."
+                ),
+                "tracking_url": "https://github.com/databricks/databricks-dbapi",


Same dead link as the ODBC spec above.

Suggested change

"tracking_url": "https://github.com/databricks/databricks-dbapi",

"tracking_url": "https://github.com/crflynn/databricks-dbapi",

Updated in ec4dc71 to point at https://github.com/crflynn/databricks-dbapi.

… dialect resolution (apache#43695) - Update tracking_url for databricks-dbapi incompatibility to https://github.com/crflynn/databricks-dbapi in DatabricksODBCEngineSpec and DatabricksHiveEngineSpec. - Verify registry.load("databricks.connector") resolves to DatabricksDialect in test_databricks_connector_dialect_registration. Signed-off-by: Francesco Castaldi <info@francescocastaldi.it>

…mat databricks test to 88 columns Signed-off-by: Francesco Castaldi <francesco.castaldi@mapsgroup.it>

Signed-off-by: Francesco Castaldi <info@francescocastaldi.it>

sadpandajoe · 2026-10-28T22:49:54Z

+        class ComplexValuesBaseEngineSpec(BaseEngineSpec):
+            engine: str = "complex_engine"
+            engine_name: str = "Complex Engine"
+            metadata: dict[str, Any] = {


The new ast.Tuple branch in _eval_ast_value has no coverage: the fixtures only use sets, negative numbers, and strings, so deleting the branch would silently turn tuple-valued metadata into None while every linter test stays green. Could a fixture add a tuple value such as "pair": ("value", -1) and assert the exact extracted tuple?

Updated ComplexValuesBaseEngineSpec fixture in tests/unit_tests/db_engine_specs/test_lint_metadata.py with "pair": ("value", -1) to explicitly assert ast.Tuple evaluation in test_get_all_engine_specs_ast_set_and_unary_minus.

FrancescoCastaldi · 2026-10-29T22:13:04Z

Summary of Review Feedback Resolutions

All inline review threads and follow-ups have been resolved and verified across the test suite:

Databricks Specifications & Dependency Alignment:
- DatabricksNativeEngineSpec: updated metadata to reference pypi_packages: ['apache-superset[databricks]'] to match modern install recommendations and avoid legacy databricks-dbapi dependency conflicts.
- DatabricksODBCEngineSpec & DatabricksHiveEngineSpec: clarified driver packages (databricks-dbapi[odbc], pyodbc and databricks-dbapi[hive], pyhive, thrift) and added explicit known_incompatibilities documenting SQLAlchemy 2.0 limitations with upstream tracking links.
- Preserved connect_args schema retention and URL synchronization in native spec.
Dialect Metadata Hardening:
- ClickHouse: explicitly configured sqlalchemy_docs_url on ClickHouseConnectEngineSpec pointing to ClickHouse Connect documentation (https://clickhouse.com/docs/en/integrations/sql-clients/superset), preventing legacy driver URL fallback through inheritance.
- IBM i: corrected the database URI parameter label to explicitly document "RDB name" rather than library/schema.
- Parseable: aligned default port 8000 with parseable+http:// while retaining parseable+https:// in connection examples.
- Aurora PostgreSQL: verified "default_port": 5432 in Data API metadata, ensuring build_sqlalchemy_uri resolves default ports without raising KeyError.
AST Linter & Test Isolation:
- Updated get_all_engine_specs_ast in lint_metadata.py to scan past bare annotations without initializers (e.g., engine: str followed by engine = "...") without prematurely stopping discovery.
- Isolated test_lint_metadata.py using pytest tmp_path fixtures and monkeypatched SPECS_DIR, completely eliminating temp file creation in the source package directory.

cc @sadpandajoe @rusackas @villebro for final review and merge.

sadpandajoe · 2026-10-29T23:01:38Z

        "logo": "clickhouse.svg",
        "logo_dark": "clickhouse-dark.svg",
        "homepage_url": "https://clickhouse.com/",
+        "sqlalchemy_docs_url": "https://clickhouse.com/docs/en/integrations/sql-clients/superset",


This sqlalchemy_docs_url redirects to clickhouse.com/docs/integrations/sql-clients/superset, which returns a 404, so the generated SQLAlchemy documentation link for ClickHouse Connect sends users to a missing page. Could it point at the live Superset guide that docs_url further down already reaches (https://clickhouse.com/docs/integrations/connectors/data-visualization/superset-and-clickhouse)?

Updated both sqlalchemy_docs_url and docs_url in ClickHouseConnectEngineSpec to the live documentation guide:
https://clickhouse.com/docs/integrations/connectors/data-visualization/superset-and-clickhouse.

sadpandajoe · 2026-10-29T23:01:38Z

+        connect_args={"schema": "finance"},
+    )
+    assert adj_url_saved.query["schema"] == "finance"
+    assert adj_args_saved["schema"] == "finance"


The or uri.query.get("schema") fallback in adjust_engine_params isn't pinned by any case: this one supplies connect_args["schema"], which wins first, and the other URI-schema cases switch catalogs or have no schema in the query. Deleting that fallback would still pass every assertion here, while /legacy?catalog=main&schema=finance with connect_args={} would then resolve to legacy. Could a case with that URI assert that both the returned URL query and connect_args carry finance?

Added test case in test_databricks_native_adjust_engine_params exercising make_url("databricks+connector://token:abc@my_hostname:1234/legacy?catalog=main&schema=finance") with empty connect_args, verifying that uri.query["schema"] is preserved and passed to both the returned URL and connect_args.

…fallback (apache#43695) - Resolve dialect catalog dynamically during get_columns and get_table_names when no catalog is configured on connection handshake. - Distinguish server default catalog from genuine catalog switches when original catalog is unconfigured. - Align ClickHouse Connect doc URLs to official live documentation guide. - Assert ast.Tuple evaluation in test_lint_metadata. - Add test coverage for catalog switch and no-catalog reflection. Signed-off-by: Francesco Castaldi <info@francescocastaldi.it>

FrancescoCastaldi · 2026-10-30T10:11:14Z

@sadpandajoe Thank you for the detailed review. All items have been addressed in the latest commit:

Databricks No-Catalog Reflection (Comment 4418792300): Avoided forcing catalog on the connection handshake (which breaks clusters without multi-catalog support) while preventing DESCRIBE TABLE EXTENDED \None`.`default`.`orders`during reflection.DatabricksNativeEngineSpec.get_columnsandget_table_namesnow resolvedialect.catalogdynamically viaSHOW CATALOGS/SELECT current_catalog()if unset prior to reflection. Addedtest_databricks_native_get_columns_resolves_no_catalog_reflection`.
Catalog Switching vs Server Default Schema Reset (Comment 4401559032): adjust_engine_params now checks catalog != server_default_catalog when no catalog is configured on the connection, ensuring genuine catalog switches reset schema to "default" while the server default retains the configured schema. Added unit test asserting genuine switches away from unconfigured catalog.
URI Schema Fallback Test Coverage (Comment 4418791833): Added unit test pinning or uri.query.get("schema") fallback in adjust_engine_params when connect_args schema is omitted.
AST Tuple Unit Test Fixture (Comment 4406914059): Added "pair": ("value", -1) fixture to tests/unit_tests/db_engine_specs/test_lint_metadata.py to assert ast.Tuple evaluation.
ClickHouse Connect Live Documentation URLs (Comment 4418791826): Updated sqlalchemy_docs_url and docs_url to the live guide https://clickhouse.com/docs/integrations/connectors/data-visualization/superset-and-clickhouse.

FrancescoCastaldi added 30 commits August 30, 2026 22:19

fix(db_engine_specs): fix AST parsing and utf-8 encoding in lint_meta…

4b8ac91

…data.py

feat(db_engine_specs): add metadata block to OdpsEngineSpec

4d91144

feat(db_engine_specs): add metadata block to IBMiEngineSpec

552c5d8

feat(db_engine_specs): add metadata block to ClickHouseEngineSpec

8114004

feat(db_engine_specs): add metadata block to OpenDistroEngineSpec

8732512

feat(db_engine_specs): add metadata block to KustoKqlEngineSpec

12f2259

feat(db_engine_specs): add metadata blocks to Databricks engine specs

87650f8

feat(db_engine_specs): add metadata blocks to Aurora engine specs

31f1108

feat(db_engine_specs): complete metadata for ArcEngineSpec

7c9873e

feat(db_engine_specs): complete metadata for ParseableEngineSpec

9c3d1ec

feat(db_engine_specs): complete metadata for OcientEngineSpec

386b94a

test(db_engine_specs): add unit tests for CloudflareD1EngineSpec

1f089cf

test(db_engine_specs): add unit tests for ArcEngineSpec

9278a0d

test(db_engine_specs): add unit tests for IBMiEngineSpec

ed47256

test(db_engine_specs): add unit tests for ParseableEngineSpec

501e0ea

test(db_engine_specs): add unit tests for OcientEngineSpec

3a1a565

test(db_engine_specs): add unit tests for OdpsEngineSpec

81214fa

test(db_engine_specs): add unit tests for HologresEngineSpec

2650d1f

test(db_engine_specs): add unit tests for AscendEngineSpec

4e09003

test(db_engine_specs): add unit tests for IoTDBEngineSpec

db65be2

test(db_engine_specs): add unit tests for ExasolEngineSpec

27fd588

test(db_engine_specs): add unit tests for MonetDbEngineSpec

6c54ec7

test(db_engine_specs): add unit tests for SolrEngineSpec

d8e6002

test(db_engine_specs): add unit tests for RisingWaveDbEngineSpec

2c4c901

test(db_engine_specs): add unit tests for PhoenixEngineSpec

ae9260d

test(db_engine_specs): add unit tests for NetezzaEngineSpec

3b7e0b2

test(db_engine_specs): add unit tests for GreenplumEngineSpec

2965c8b

test(db_engine_specs): add unit tests for YugabyteDBEngineSpec

2be71ba

test(db_engine_specs): add unit tests for ShillelaghEngineSpec

8b9e2cd

test(db_engine_specs): add unit tests for TeradataEngineSpec

bf8b5cb

github-actions Bot removed the requires:rebase Requires rebasing on top of current master label Oct 13, 2026

sadpandajoe reviewed Oct 14, 2026

View reviewed changes

Comment thread superset/db_engine_specs/databricks.py Outdated

Comment thread tests/unit_tests/db_engine_specs/test_databricks.py Outdated

sadpandajoe reviewed Oct 14, 2026

View reviewed changes

Comment thread superset/db_engine_specs/databricks.py

Comment thread superset/db_engine_specs/lint_metadata.py

sadpandajoe reviewed Oct 14, 2026

View reviewed changes

Comment thread superset/db_engine_specs/databricks.py Outdated

Comment thread superset/db_engine_specs/databricks.py Outdated

Comment thread tests/unit_tests/db_engine_specs/test_lint_metadata.py Outdated

sadpandajoe reviewed Oct 15, 2026

View reviewed changes

Comment thread tests/unit_tests/db_engine_specs/test_databricks.py

Comment thread tests/unit_tests/db_engine_specs/test_databricks.py

Comment thread tests/unit_tests/db_engine_specs/test_lint_metadata.py

sadpandajoe approved these changes Oct 15, 2026

View reviewed changes

Merge branch 'master' into feat/db-engine-specs-improvements

e3b3fad

sadpandajoe reviewed Oct 21, 2026

View reviewed changes

Comment thread superset/db_engine_specs/databricks.py Outdated

Comment thread tests/unit_tests/db_engine_specs/test_databricks.py

rusackas reviewed Oct 27, 2026

View reviewed changes

sadpandajoe reviewed Oct 28, 2026

View reviewed changes

Comment thread tests/unit_tests/db_engine_specs/test_databricks.py Outdated

Comment thread tests/unit_tests/db_engine_specs/test_databricks.py

FrancescoCastaldi added 2 commits October 28, 2026 15:20

test(db_engine_specs): exercise connect_args schema retention and for…

9e65e1d

…mat databricks test to 88 columns Signed-off-by: Francesco Castaldi <francesco.castaldi@mapsgroup.it>

style(databricks): break line 272 comment to satisfy 88-char ruff limit

d511920

Signed-off-by: Francesco Castaldi <info@francescocastaldi.it>

sadpandajoe reviewed Oct 28, 2026

View reviewed changes

sadpandajoe reviewed Oct 29, 2026

View reviewed changes

	"tracking_url": "https://github.com/databricks/databricks-dbapi",
	"tracking_url": "https://github.com/crflynn/databricks-dbapi",

Conversation

FrancescoCastaldi commented Aug 30, 2026

Summary

Key Changes

Testing Instructions

Uh oh!

FrancescoCastaldi commented Oct 14, 2026

Uh oh!

Uh oh!

Uh oh!

FrancescoCastaldi commented Oct 14, 2026

Uh oh!

Uh oh!

Uh oh!

FrancescoCastaldi commented Oct 14, 2026

PR Update: Databricks Default Schema Resolution and CLI ASCII Stdout Test Coverage

Changes Applied (Commit 1a59b994)

Uh oh!

Uh oh!

Uh oh!

Uh oh!

FrancescoCastaldi commented Oct 14, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

FrancescoCastaldi commented Oct 15, 2026

Uh oh!

Uh oh!

Uh oh!

rusackas commented Oct 21, 2026

Uh oh!

FrancescoCastaldi commented Oct 24, 2026

Uh oh!

rusackas left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

FrancescoCastaldi commented Oct 29, 2026

Summary of Review Feedback Resolutions

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

FrancescoCastaldi commented Oct 30, 2026

Uh oh!

This branch has not been deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Changes Applied (Commit `1a59b994`)