schema-validator

name	schema-validator
description	Validate tabular data against a predefined JSON schema — checks column presence, data types, nullable constraints, and unique constraints.
version	0.6.0

Schema Validator

Skill Summary

Validates a tabular dataset against a user-defined JSON schema. Checks that all required columns are present, column data types match expectations, nullable constraints are respected, and unique constraints hold. Produces a structured validation report indicating pass/fail for each rule, with row-level error details.

Schema Definition Format

{
  "columns": {
    "column_name": {
      "type": "string|int|float|date|bool",
      "nullable": false,
      "unique": false,
      "pattern": "regex (for string columns)"
    }
  },
  "required_columns": ["col1", "col2"]
}

Inputs

• --input / -i: Path to dataset file (CSV or Parquet) (required)
• --schema / -s: Path to JSON schema file (required)
• --output / -o: Output path for validation report
• --strict: Fail if extra columns exist beyond schema (default: false)
• --max-errors: Stop after N errors (default: unlimited)

Validation Checks

1. Required columns: All columns listed in required_columns must exist
2. Type checking: Each column value must match declared type
3. Nullable: Non-nullable columns must have zero null values
4. Uniqueness: Unique columns must have no duplicate values
5. Pattern matching: String values must match declared regex pattern
6. Extra columns: In strict mode, flag columns not in schema

Output

• validation_report.json: Structured report with:
  • Overall pass/fail status
  • Per-column validation results
  • Row-level error details (row index, column, expected vs actual)
  • Error counts by type

Implementation

• Command: python3 ./skills/schema_validator/validate.py -i <file> -s <schema>
• Dependencies: pandas, jsonschema