A stakeholder objects that object storage cannot be used for a SQL platform because SQL requires a database. What is the correct response?
Select an answer to reveal the explanation.
Short Explanation
SQL needs a table, not a database server. The catalog supplies the schema, the table format supplies the semantics, and the engine executes the query while the bytes stay in the bucket. That layering is the whole lakehouse idea.
Full Explanation
The correct response is that engines query open table formats defined over objects through catalog metadata, so SQL works against object storage without a database engine owning it. SQL requires structure and an execution engine, both of which the lakehouse supplies without a database owning the storage. The catalog defines tables over sets of objects, open table formats give those tables transactional and schema semantics, and query engines plan and execute against them.
The response saying the objection is valid because watsonx.data loads object data into an internal database before queries run is incorrect because there is no hidden load step into an internal database. Engines read the data files directly; that absence of a copy is a defining property of the architecture.
The response saying the objection is valid for reads but not for writes, which go directly to objects, is incorrect because reads and writes both operate through the same table abstraction. Writes append new data files and commit new metadata rather than bypassing the model.
The response saying SQL runs against object storage only when the objects are stored as CSV is incorrect because file format is not what makes SQL possible, and CSV is in fact a poor choice - it carries no schema evolution, no statistics and no transactional guarantees.
For the exam, be able to state the stack in order: objects, open table format, catalog metadata, engine.