Feather

提供 Apache Feather/Arrow IPC 列式格式的文档参考，覆盖 V1/V2 差异、pyarrow.feather 读写与 LZ4/ZSTD 压缩、R arrow 包互操作、Arrow 类型系统、与 Parquet 的速度与压缩权衡、pandas 缓存模式及流水线最佳实践；通过命令选择主题并输出对应说明文本。

bytesagain1

热度 51.70

最近更新 2026-05-11 17:35:51

分类：知识依赖：代码执行应用场景：概念解释 / 数据处理

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add feather

Metadata

name: feather
version: 1.0.0
description: Apache Feather/Arrow IPC format reference. V1 vs V2 format differences, pyarrow.feather read/write with compression, R arrow package integration, Arrow type system, Feather vs Parquet benchmarks, pandas DataFrame caching, LZ4/ZSTD compression options, and pipeline best practices.
author: BytesAgain
homepage: https://bytesagain.com
source: https://github.com/bytesagain/ai-skills
tags: ["feather","arrow","ipc","columnar","pandas","dataframe","data"]
category: data

Feather

Apache Feather/Arrow IPC format reference — fast columnar DataFrame I/O.

Commands

Command	Description
`intro`	Feather overview, V1 vs V2, key properties
`python`	pyarrow.feather read/write, benchmarks
`r-lang`	R arrow package, Python↔R interop
`schema`	Arrow types, nested types, metadata
`vs-parquet`	Speed vs compression tradeoffs
`pandas`	pd.read_feather, caching patterns
`compression`	LZ4/ZSTD options, when to use each
`best-practices`	Pipeline patterns, file size guidelines