Expose Druid functions in INFORMATION_SCHEMA.ROUTINES table.

[abhishekrb19]

This PR adds a new table INFORMATION_SCHEMA.ROUTINES. Resolves #14346.

Description

The implementation currently exposes information about all functions in INFORMATION_SCHEMA.ROUTINES. The table contains the following columns:

• ROUTINE_CATALOG: Name of the database. Always set to druid.
• ROUTINE_SCHEMA: Name of the schema. Always set to INFORMATION_SCHEMA.
• ROUTINE_NAME: Name of the function or operator. E.g., APPROX_COUNT_DISTINCT_DS_THETA
• ROUTINE_TYPE: Type of routine. Always set to FUNCTION.
• IS_AGGREGATOR: Is the routine an aggregator or not. Returns YES for aggregator functions; NO for scalar functions.
• SIGNATURES: Possible signatures for the routine as a string.

Since Calcite is the source of truth for all the registered functions, including ones loaded through custom extensions, the implementation uses the Calcite's operator table to extract this information at runtime.

A preview of the table from the Druid console:

Release note

Expose Druid functions via INFORMATION_SCHEMA.ROUTINES table.

---

Key changed/added classes in this PR

• InformationSchema.java
• DruidOperatorTable.java
• InformationSchemaTest.java
• DruidOperatorTableTest.java

---

This PR has:

• been self-reviewed.
  • using the concurrency checklist (Remove this item if the PR doesn't have any relation to concurrency.)
• added documentation for new or modified features or behaviors.
• a release note entry in the PR description.
• added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
• added or updated version, license, or notice information in licenses.yaml
• added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.
• been tested in a test Druid cluster.

[FrankChen021]

I'm wondering why is not the IS_AGGREGATOR declared as BOOLEAN?

IMO for boolean type, it would be more simplier and intuitive for user to filter out aggregator functions like

select * from ROUTINES where IS_AGGREGATOR

[abhishekrb19]

I'm wondering why is not the IS_AGGREGATOR declared as BOOLEAN?

IMO for boolean type, it would be more simplier and intuitive for user to filter out aggregator functions like

select * from ROUTINES where IS_AGGREGATOR

Good question, @FrankChen021. I wondered the same :)
The INFORMATION_SCHEMA tables in Druid already have boolean values converted to varchar - for example, IS_JOINABLE, IS_BROADCAST, IS_NULLABLE. The SQL spec notes that, so most standard implementations are just following the spec. So for consistency, I made the new column VARCHAR, as it doesn't deviate too much from the spec.

[FrankChen021]

OK, let's keep it consistent with existing designs.

[zachjsh]

🚀