gate schema discovery behind "auto" for remote engines

[Light2Dark]

📝 Summary

Closes #9775

The datasources panel scanned every catalog/schema on remote backends such as Snowflake because the ibis and adbc engines hardcoded auto_discover_schemas=True. Switch them to "auto" so discovery is gated by the cheap-dialect heuristic, matching the sqlalchemy engine.

Also unify the shared default discovery config (default_inference_config) and the "auto" resolution (_resolve_should_auto_discover / _is_cheap_discovery) across engines to prevent this kind of drift in the future.

📋 Pre-Review Checklist

• For large changes, or changes that affect the public API: this change was discussed or approved through an issue, on Discord, or the community discussions (Please provide a link if applicable).
• Any AI generated code has been reviewed line-by-line by the human PR author, who stands by it.
• Video or media evidence is provided for any visual changes (optional).

✅ Merge Checklist

• I have read the contributor guidelines.
• Documentation has been updated where applicable, including docstrings for API changes.
• Tests have been added for the changes made.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
marimo-docs	Ready	Preview, Comment	Jun 15, 2026 3:39am

[Copilot]

Pull request overview

This PR fixes a performance issue in the datasources panel for remote SQL backends (notably Snowflake) by ensuring schema discovery is no longer unconditionally enabled for the Ibis and ADBC engines. It aligns discovery behavior across engines by centralizing the default inference config and the "auto" discovery resolution heuristic, matching the existing intent of the SQLAlchemy engine.

Changes:

• Switch Ibis and ADBC engines from auto_discover_schemas=True to a shared default_inference_config() where schemas/tables are gated by "auto".
• Centralize "auto" resolution and cheap-dialect detection via EngineCatalog._resolve_should_auto_discover() / _is_cheap_discovery() and is_cheap_dialect().
• Add regression tests asserting SQLAlchemy/Ibis/ADBC engines use the shared default inference config (and that schemas default to "auto").

Reviewed changes

Copilot reviewed 11 out of 11 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
marimo/_sql/utils.py	Adds is_cheap_dialect() helper to centralize the “cheap discovery” dialect heuristic.
marimo/_sql/engines/types.py	Introduces default_inference_config() and centralizes "auto" discovery resolution in the base engine catalog type.
marimo/_sql/engines/sqlalchemy.py	Switches SQLAlchemy engine to use default_inference_config() and removes duplicated "auto" resolution logic.
marimo/_sql/engines/ibis.py	Fixes Ibis schema discovery default by using default_inference_config() (prevents eager schema scanning on remote backends).
marimo/_sql/engines/adbc.py	Fixes ADBC schema discovery default by using default_inference_config(); uses shared cheap-dialect helper.
marimo/_sql/engines/redshift.py	Updates engine override to implement _is_cheap_discovery() instead of duplicating "auto" resolution.
marimo/_sql/engines/pyiceberg.py	Removes duplicated "auto" resolver override; relies on base resolver + _is_cheap_discovery() override.
marimo/_sql/engines/clickhouse.py	Updates engine override to implement _is_cheap_discovery() instead of duplicating "auto" resolution.
tests/_sql/test_sqlalchemy.py	Adds test asserting SQLAlchemy engine uses the shared default inference config.
tests/_sql/test_ibis.py	Adds test asserting Ibis engine uses the shared default inference config and schemas default to "auto".
tests/_sql/test_adbc.py	Adds test asserting ADBC engine uses the shared default inference config and schemas default to "auto".

[cubic-dev-ai]

4 issues found across 11 files

Architecture diagram

sequenceDiagram
    participant UI as Datasources Panel
    participant Engine as SQL Engine (base)
    participant Adbc as ADBC Engine
    participant Ibis as Ibis Engine
    participant SA as SQLAlchemy Engine
    participant ClickHouse as ClickHouse Engine
    participant Redshift as Redshift Engine
    participant PyIceberg as PyIceberg Engine
    participant Types as types.py
    participant Utils as utils.py
    
    Note over UI,PyIceberg: Schema/Table Discovery Flow
    
    UI->>Engine: request datasource discovery
    Engine->>Engine: inference_config
    Note over Engine: Default: auto_discover_schemas="auto"
    
    alt Engine is ADBC or Ibis or SQLAlchemy
        Engine->>Types: call default_inference_config()
        Types-->>Engine: InferenceConfig with auto_discover_schemas="auto"
    else Engine is ClickHouse or Redshift
        Engine->>Engine: hardcoded override (not using default)
        Note over Engine: _is_cheap_discovery() returns False
    else Engine is PyIceberg
        Engine->>Engine: hardcoded override
        Note over Engine: _is_cheap_discovery() returns True
    end
    
    Engine->>Engine: _resolve_should_auto_discover("auto")
    alt Resolution by dialect heuristic
        Engine->>Utils: is_cheap_dialect(dialect)
        Utils->>Utils: check CHEAP_DISCOVERY_DATABASES set
        alt dialect in cheap set (e.g., sqlite, duckdb)
            Utils-->>Engine: return True
            Engine-->>UI: auto-discover schemas
        else dialect not in cheap set (e.g., snowflake, bigquery)
            Utils-->>Engine: return False
            Engine-->>UI: skip auto-discovery
        end
    else Override by engine policy
        ClickHouse-->>ClickHouse: return False (always expensive)
        Redshift-->>Redshift: return False (opt out)
        PyIceberg-->>PyIceberg: return True (metadata always fast)
    end
    
    Note over UI,PyIceberg: Dialect-based gating unified via base class

Loading

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[Copilot]

Pull request overview

Copilot reviewed 12 out of 12 changed files in this pull request and generated no new comments.

[mscolnick]

could this be a set? not sure where it is used elsewhere