Description

Preserve an explicitly configured BigQuery dataset ID when that exact dataset already exists.

The warehouse namespace builder snake-cases custom namespaces before the BigQuery manager is created. bq_events_zm_dev_dts_euwe1, for example, becomes bq_events_zm_dev_dts_euwe_1 due to its letter-to-digit boundary. The BigQuery setup path now checks the original configured dataset after establishing its client. If it exists, subsequent schema management and load operations use it verbatim. If it does not exist, the existing sanitised namespace behaviour remains unchanged.

Added regression coverage for the exact letter/digit case, absent datasets, both datasets existing, and failed/no-op checks.

Linear Ticket

No linear provided, but I have filed a support ticket for a finding, this would solve that (I think)

Security

• The code changed/added as part of this pull request won't create any security issues with how the software is being used.

New behavior

• Exact configured dataset exists: preserve its ID, including letter-to-digit boundaries.
• Exact configured dataset is absent: retain the existing sanitised name and creation flow.
• Dataset lookup errors: retain the existing sanitised name rather than failing a previously valid setup path.