FrameX Documentation

Features

Core FrameX capabilities across DataFrame, NDArray, runtime, and interoperability.

Features

FrameX focuses on high-throughput local analytics with predictable behavior and practical interoperability.

DataFrame Features

  • Arrow-backed partitioned DataFrame and Series
  • Filtering, projection, sorting, joins, and groupby aggregation
  • Eager API with optional lazy pipelines (.lazy().collect())
  • Pandas-compatible fallback for unimplemented methods

NDArray Features

  • Chunked NDArray with NumPy protocol support:
    • __array_ufunc__
    • __array_function__
  • Block-parallel operations:
    • .apply_blocks(...)
    • .parallel_map(...)
    • .jit_apply(...)

Runtime Features

  • Local backends: threads, processes
  • Optional distributed backends: ray, dask, hpc
  • Hardware-aware auto configuration:
    • recommend_best_performance_config()
    • auto_configure_hardware()

Interoperability Features

  • from_pandas(...), from_dask(...), from_ray(...)
  • .to_pandas(), .to_arrow(), .to_dask(), .to_ray()
  • DataFrame interchange protocol support

I/O Features

  • Unified read_file(...) / write_file(...)
  • Formats:
    • Parquet, ORC, Arrow IPC
    • CSV/TSV/Text + fixed-width text
    • JSON/NDJSON, Feather, Pickle, Excel, SQLite
    • Export-only: HTML and XML
  • Compression wrappers:
    • .gz, .bz2, .xz, .zip
    • .zst/.zstd (with zstandard)