Using Spider2-Snow instead of Spider2-Lite due to BigQuery service-account issue

[RomanAvdeev]

Hello Spider 2.0 team,

I am currently trying to run my pipeline on Spider2-Lite. The SQLite part works, and I have already requested Snowflake access from your team.

However, I am blocked on the BigQuery side.

What I have already checked:

• created a Google Cloud project.
• created a service account and generated a valid JSON key.
• enabled BigQuery API.
• granted the service account project-level roles including:
  • BigQuery Job User
  • BigQuery User
  • BigQuery Admin
  • Owner
• tested the credential using the exact Python access pattern shown in your BigQuery guideline

What happens:

• The public example query works in the BigQuery web UI under my user account.
• But every API-based query from the service account fails with:

403 Forbidden: POST https://bigquery.googleapis.com/bigquery/v2/projects/<project_id>/jobs

So at the moment, the issue seems not to be the JSON key format or the client code, but that BigQuery jobs cannot be created through the service account in this project.

Because of this, I am considering switching from Spider2-Lite to Spider2-Snow once I receive Snowflake credentials from your team.

Could you please confirm:

1. After receiving Snowflake access, will Spider2-Snow provide full access to the complete 547-task benchmark through Snowflake alone?
2. Is Spider2-Snow a valid substitute for Spider2-Lite if BigQuery service-account access cannot be made to work?

Thank you very much for your help.