dataframe-core
Create, inspect, filter, and validate Vaex DataFrames while preserving lazy out-of-core semantics.
- License 许可
- MIT
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- dataframe-core
- license
- MIT
- metadata
- { "disco-role": "operating" }
- description
- Create, inspect, filter, and validate Vaex DataFrames while preserving lazy out-of-core semantics.
- disable-model-invocation
- true
02 资源文件 4
- 参考文档references/api-reference.md8.7 KB
- 参考文档references/dataframe-workflows.md8.9 KB
- 参考文档references/troubleshooting.md6.8 KB
- 脚本scripts/dataframe_smoke.py3.5 KB
03 SKILL.md 正文
dataframe-core
Use this sub-skill when a task is about the Vaex DataFrame object itself: creating small or large DataFrames, inspecting columns and shape, writing filters or selections, adding virtual columns, validating expressions, and debugging missing values or unusual column names without accidentally materializing an out-of-core dataset.
Load these references
- Start with references/dataframe-workflows.md for concrete recipes: constructors, inspection, slicing/filtering, selections, virtual columns, non-identifier column names, missing values, and Pandas-to-Vaex translations.
- Use references/api-reference.md for verified signatures and object-method notes.
- Use references/troubleshooting.md when imports, file opening, expressions, filters, missing values, or memory use surprise the user.
- Run scripts/dataframe_smoke.py as a safe installed-package check before relying on Vaex core behavior in a new environment.
Core operating rules
- Treat Vaex as lazy and out-of-core by default. Prefer expressions, virtual columns, selections, filters,
count, and aggregations over.values,.to_numpy(),np.array(df),to_pandas_df, or fullevaluatecalls. - Use
evaluateintentionally and bounded: passi1/i2,selection=...,array_type=..., or iterate with chunks when the result may be large. - For column names that contain spaces, punctuation, keywords, or symbols, use bracket access (
df['column name']) or pass theExpressionobject to methods instead of relying on attribute access or raw expression strings. - Distinguish filters from selections.
df[df.x > 0]anddf.filter(...)return shallow filtered DataFrames;df.select(...)stores a named selection on the same DataFrame for statistics, evaluation, and visualization. - Add derived columns as virtual columns (
df['new'] = expressionordf.add_virtual_column(...)) unless an in-memory array is intentionally small and has the unfiltered DataFrame length.
Boundaries
- Route detailed file formats, export, conversion, cloud/open plugin setup, and CLI conversion tasks to
../io-conversion/SKILL.md. - Route statistical, groupby, binby/grid, join, sort, unique, and value-count recipes to
../expressions-analytics/SKILL.md; this sub-skill only introduces the concepts needed to inspect or validate the DataFrame. - Route ML transformations,
vaex.ml, pipelines, encoders, scalers, and sklearn integration to../ml-pipelines/SKILL.md. - Route plotting, Jupyter widgets, progress-bar UI, and visualization troubleshooting to
../visualization-jupyter/SKILL.md.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).