Full-text search returns nothing instead of scanning raw data when no full-text index exists

[LuciferYang]

Search before asking

• I searched in the issues and found no similar issues.

Paimon version

master (1.5-SNAPSHOT)

Compute Engine

Spark / Flink full-text search over a data-evolution table (with paimon-full-text).

Minimal reproduce step

1. Create a data-evolution table with a full-text-indexed column.
2. Write data but do not build the full-text index (index building is a separate commit from data writes), or query after the index has expired.
3. Run a full-text search in FULL or DETAIL mode.

What doesn't meet your expectations?

The search returns zero rows instead of scanning the raw data. In DataEvolutionFullTextScan, the raw-data compensation split (for rows not covered by a full-text index) is gated on if (!fullTextIndexFiles.isEmpty()), so when no index exists the compensation is skipped and the plan is empty. unindexedRanges would correctly return the whole row-id space (FULL) or the data-file ranges (DETAIL). FAST mode is correctly empty.

Anything else?

There is also a read-side checkNotNull on a null index type that would NPE once the scan-side gate is removed; the raw fallback index type needs to resolve to the built-in full-text type.

Are you willing to submit a PR?

• I'm willing to submit a PR!