菜单
Use DuckDB to query Parquet files, inspect metadata, join tables, and convert results to pandas for analysis; commonly precedes ydata-eda-profiling for EDA on extracted tables.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/crossxwill/IML4Finance --skill duckdb-parquet-lab-workflow命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Edit Quarto (.qmd) and Markdown (.md) files with proper nested code block handling to prevent parser errors and truncated output; standalone utility with no skill dependencies.
Render or build or compile Quarto .qmd files from the command line.
Consolidated AutoGluon skill covering end-to-end TabularPredictor workflow (constructor, fit, predict_proba, fit_summary, save/load, set_model_best), binary threshold calibration and setting, sklearn wrapper integration, and monotonic constraints. Use for any AutoGluon Tabular questions, training/ensembling configuration, inference/probabilities, threshold tuning, deployment persistence, or sklearn interoperability.
基于 SOC 职业分类
正在显示 SKILL.md
| name | duckdb-parquet-lab-workflow |
| description | Use DuckDB to query Parquet files, inspect metadata, join tables, and convert results to pandas for analysis; commonly precedes ydata-eda-profiling for EDA on extracted tables. |
Standardize the pattern of loading Parquet files into DuckDB, inspecting schema, running SQL joins, and converting results to pandas DataFrames.
duckdb.query or duckdb.sql using SQL strings.DESCRIBE SELECT * FROM <table> and display with .show().LEFT or RIGHT semantics to preserve row counts..to_df() for downstream modeling../templates/duckdb_snippets.md for the standard SQL patterns.