Cannot parse as CloudRegion

[0anton]

For every request, for example this:

{
  `sql`: `
SELECT schema_name
FROM INFORMATION_SCHEMA.SCHEMATA
ORDER BY schema_name;
`
}

I'm getting an error:

Error executing code: MCP error 400: Cannot parse as CloudRegion.

There is a double space after the word "parse". I don't see the word CloudRegion in the source code of this module.

I've tried both project id and project number - same error.

bq cli can list table with out any problems.

[silvanm]

In my case, the problem was the multi-region setup of my DB. I modified the code a bit to accommodate for this.

#5

[0anton]

Thanks für PR, @silvanm! The fixed list of regions is a bit risky (not „timeless“), as Google adds regions from time to time, but it is easy to correct. I like the idea with default region „EU“, too 🙂, but sure our US friends will disagree with its commonality.

I‘ve noticed the BigQuery MCPs all require the region specified. It‘s very annoying if you have tables in mixed regions (e.g. „EU“ and „europe-west1“), as it randomly fails as traffic moves from table to table.

If I run „bq query“ from CLI I never have to bother about region. It just runs, as soon as I have permissions to the object. I wonder if MCP can also be made region agnostic….

I also don‘t remember if BigQuery API requires regions specification.

[ngohuuduc]

@0anton Do you still face the same issue ? I'm facing this issue and not able to solve.

[0anton]

@ngohuuduc yes

[ngohuuduc]

@ngohuuduc yes

seems like the issue is actually coming from API itself ?

[ngohuuduc]

@0anton : you may try my fork version
https://github.com/ngohuuduc/mcp-bigquery-server

it's just work for me.