feat(bigquery-jdbc): support qualified project delimiter in DefaultDataset property

[keshavdandeva]

b/556170877

Parse DefaultDataset with both ':' and '.' delimiters to support PCNT datasets and projects

Key Changes

• Unified DefaultDataset Parsing: Extracted BigQueryJdbcUrlUtility.parseDefaultDataset() to support both : (e.g. project:warehouse.namespace) and . delimiters, ensuring consistent parsing across JDBC connection strings and DataSource configurations.
• Metadata Filtering: Updated BigQueryDatabaseMetaData.determineEffectiveCatalogAndSchema() to use defaultDataset.getProject() when FilterTablesOnDefaultDataset=1 is enabled.
• Catalog Resolution: Updated BigQueryDatabaseMetaData.getAccessibleCatalogNames() to include the project from defaultDataset in the list of accessible catalogs.
• Testing: Added focused unit tests in BigQueryJdbcUrlUtilityTest and BigQueryDatabaseMetaDataTest covering multi-format parsing, namespace delimiters, and catalog filtering.

[gemini-code-assist]

Code Review

This pull request refactors the default dataset parsing logic in BigQueryConnection by delegating it to a new utility method BigQueryJdbcUrlUtility.parseDefaultDataset, which supports both colon- and dot-delimited formats. It also updates BigQueryDatabaseMetaData to correctly resolve catalogs and schemas when a default dataset project is specified, and adds comprehensive unit tests. The review feedback suggests using lastIndexOf(':') instead of indexOf(':') to properly support domain-scoped projects (e.g., example.com:project-id:dataset-id) and recommends adding test cases to cover this scenario.

[keshavdandeva]

/gemini review

[gemini-code-assist]

Code Review

This pull request refactors default dataset parsing by delegating it to a new utility method BigQueryJdbcUrlUtility.parseDefaultDataset and updates BigQueryDatabaseMetaData to correctly handle default datasets that include a project ID. It also adds corresponding unit tests. The review feedback highlights a bug in the new parsing logic where domain-scoped projects using a dot separator (e.g., example.com:my-project.my_dataset) are parsed incorrectly, and provides a robust implementation to resolve this along with a suggestion to add a test case for this scenario.