profiling-tables

Table profiling gathers schema, size, and statistical summaries to help you understand structure and content. It collects column metadata, row counts, null/empty counts, distinct counts, and representative samples to reveal data distribution and anomalies for analytics or ETL decisions.

Data profiling computes row counts, nulls, empties, distinct counts, min/max/avg for numeric columns, and date ranges to reveal shape and anomalies. It also identifies skewed categorical values through cardinality analysis and generates representative samples to speed up exploratory work.

Yes, DuckDB supports SQL-based metadata queries over INFORMATION_SCHEMA needed for table profiling. The skill applies these queries plus aggregate computations for numeric, string, and date/timestamp columns to gather schema, size, and statistical summaries for data discovery and pipeline debugging.

Assessing data quality requires computing null/empty counts, distinct counts, and cardinality to identify anomalies and skewed values. The skill produces an interpretable summary including an overall quality scoring rubric and recommended follow-up queries to guide your analytics or ETL decisions.

The best way to onboard is running schema-first profiling that collects column names, types, and comments alongside distribution statistics. The skill generates a decision-ready profile with representative samples and recommended follow-up queries, helping you quickly understand structure and content for integration.