[lucasfang]

Purpose

Add regression coverage for late materialization and Parquet page-index filtering, and clarify the README's build requirements and example entry points.

• Cover payload early EOF, empty payload selections, and batches with no selected rows.
• Cover IN/NOT IN with null literals, missing null counts, and truncated integer bounds.
• Link the build guide and both standalone examples from the README.

Tests

• pre-commit run --files README.md src/paimon/common/reader/late_materializing_file_batch_reader_test.cpp src/paimon/format/parquet/column_index_filter_test.cpp — passed.
• git diff --check and git diff --cached --check — passed before committing.
• Rebuilt paimon-common-test and paimon-parquet-format-test successfully with PATH=/home/yonghao.fyh/.cargo/bin:/home/yonghao.fyh/tools/cmake-3.30.0/bin:$PATH RUSTUP_TOOLCHAIN=1.94.0 cmake --build build --target paimon-common-test paimon-parquet-format-test -j 8.
• On the rebuilt binaries, LateMaterializingFileBatchReaderTest.* passed 25 tests and ColumnIndexFilterTest.* passed 24 tests through build_support/run-test.sh, including all six added cases.
• Both standalone example targets, read_write_demo and clean_demo, compiled and linked using GCC 10.2.1, C++17, Arrow 17 and the existing Debug Paimon installation. Examples were not executed.

The initial test-target rebuild encountered the system Rust 1.67.1 toolchain; using the installed Rust 1.94 toolchain resolved that environment issue. Full unit and integration suites have not been run.

API and Format

No production code, public API, storage format, or protocol changes.

Documentation

Document compiler/CMake requirements, link the build guide, and add read/write and cleanup example links. Adjust feature terminology for consistency.

Generative AI tooling

Generated-by: OpenAI Codex (GPT-6)

[lxy-9602]

+1