Why is a catalog necessary in a lakehouse when the data files themselves already contain schema information?
Select an answer to reveal the explanation.
Short Explanation
Parquet files know their own columns, but they don’t know they’re part of a table, which snapshot is current, or where to find the whole set. You need the catalog to name the table, track its files, and make it discoverable for queries.
Full Explanation
The correct explanation is that file-level schema does not say which files form a table, what its current snapshot is, or where to find it; the catalog supplies that discoverable, authoritative view. Self-describing files carry their own column definitions but nothing about membership or identity. The catalog names tables, records which files constitute each snapshot, and makes them discoverable so an engine can resolve a table reference in SQL without scanning a bucket to guess.
The claim that data files do not actually contain schema information, so the catalog is the only source, is incorrect factually because Parquet embeds its schema. The catalog's role is table-level, not column-level rescue.
The claim that the catalog is needed only for engines that cannot read Parquet natively is incorrect because engines that read Parquet perfectly well still need to know which files make up a given table at a given moment.
The claim that the catalog exists to compress schema information and reduce storage cost is incorrect because storage savings on schema metadata are negligible and are not why catalogs exist.
For the exam, phrase the catalog's role as identity, membership and discoverability of tables.