Profile Dataset
AI-generatedSummary
Profile a JSON array of objects representing a dataset, analyzing each column for data types, null values, uniqueness, numeric statistics, duplicates, outliers, and constant columns, returning a comprehensive profiling report including warnings and audits.
Inputs
- Treat Empty Strings as Null — Optional boolean to treat blank string values as nulls for data quality metrics; defaults to true.
- Coerce Numeric Strings — Optional boolean to interpret numeric-looking strings as numbers during profiling; defaults to true.
Output shape
A single object containing the dataset profile report, including row and column counts, array of column profiles with type, nulls, cardinality, numeric stats if applicable, detected duplicate rows with groups, outliers detected via the interquartile range method, constant columns, and summary statistics. Also includes metadata with timing and counts, any warnings (duplicate rows, outliers, constant columns, high null ratio), errors (e.g., input validation), and an audit trail of processing steps.
The input must be an array of JSON objects, each representing a dataset row. Profiling respects the two boolean options to affect null value treatment and numeric coercion. Results include potential warnings about data quality issues. If input validation fails, an error result is returned instead.
Examples
Example 1: Profiling a JSON dataset with default settings treating empty strings as null and coercing numeric strings.
treatEmptyStringAsNull: true, coerceNumericStrings: true
Example 2: Profiling a dataset where numeric strings should not be coerced to numbers.
treatEmptyStringAsNull: true, coerceNumericStrings: false