Correctly handle double dashes in BigQuery project names

[erindru]

Currently, trying to specify a BigQuery project with dashes in it just produces errors trying to fetch column info.

Against our sqlmesh init example structure, with the following config:

gateways:
  bigquery:
    connection:
      type: bigquery
      project: "`arcx--dev`" #or project: "arcx--dev"

BigQuery produces errors like:

google.api_core.exceptions.Forbidden: 403 Access Denied: Table arcx:sqlmesh__sqlmesh_example.INFORMATION_SCHEMA.COLUMNS: User does not have permission to query table arcx:sqlmesh__sqlmesh_example.INFORMATION_SCHEMA.COLUMNS, or perhaps it does not exist.;

(notice SQLMesh tried arcx:sqlmesh__sqlmesh_example instead of arcx--dev:sqlmesh__sqlmesh_example)

Also, the proper catalog is clearly missing from the model names in the plan:

[georgesittas]

I don't think using backticks there is the right syntax. Both project: foo-bar and project: "foo-bar" work as is for me. I'll go ahead and close this as a non-issue, but let me know if I missed something.

[tanghyd]

I don't think using backticks there is the right syntax. Both project: foo-bar and project: "foo-bar" work as is for me. I'll go ahead and close this as a non-issue, but let me know if I missed something.

Have you tried project: "foo--bar"? The issue was specifically about double dashes.

[erindru]

@georgesittas -- is a SQL comment, right? I bet we aren't handling this properly somewhere.

For example, using exp.to_table without quoting correctly:

>>> from sqlglot import exp
>>> exp.to_table("foo--bar.baz.bing")
Table(
  this=Identifier(this=foo, quoted=False),
  _comments=[
    bar.baz.bing])

[georgesittas]

Ah nice catch. I definitely missed this :)
Will fix shortly.