AgentSkillsCN

data-sanity-check

对数据框或数组进行模式、空值、取值范围及重复值的校验。

SKILL.md
--- frontmatter
name: data-sanity-check
description: Validate dataframes or arrays for schema, nulls, ranges, and duplicates.

Data Sanity Check

When to use

Use after data loading, feature generation, or merging datasets.

Inputs

  • •DataFrame(s) or array(s)
  • •Expected schema (optional)

Steps

  1. •Report shapes and dtypes.
  2. •Check null counts and percentages.
  3. •Validate value ranges for critical columns.
  4. •Detect duplicates on key columns.
  5. •Flag unexpected columns or missing required fields.

Outputs

  • •Summary table (shape, dtypes)
  • •Anomaly report (nulls, range violations, duplicates)

Failure conditions

  • •Critical nulls or schema mismatch
  • •Out-of-range values beyond tolerance