liveBook · Manning

Parquet_metadata

As explained in DuckDB in Action

parquet_metadata is a DuckDB table function that returns metadata for each row group and column in a Parquet file.

parquet_metadata is a DuckDB table function for inspecting Parquet structure and statistics. It returns one record for each row group and column, including types, sizes, compression, null counts, and value bounds. section 5.5

Inspect the first row-group column

Run `FROM parquet_metadata('atp/atp_rankings.parquet') LIMIT 1;` in line mode to inspect a compact record. For the ranking_date column in row group 0, stats_min is 1973-08-27 and stats_max is 1979-12-26. DuckDB can use these bounds to skip the row group when a query requests dates after 1980. section 5.5

How DuckDB in Action covers it

How deep it goes
Explains row-group and column records with statistics, types, sizes, and compression, but stops short of broader Parquet internals.
Examples
Uses an ATP rankings Parquet file and examines ranking_date bounds in its first row group.
What it assumes
Assumes basic DuckDB SQL and Parquet familiarity while demonstrating file inspection directly without persistence.
What you can do after
Evaluate Parquet layout and statistics, then understand when DuckDB can safely skip row groups during filtering.
How it connects
Connects Parquet files, row groups, column statistics, data types, compression, and query optimization.
Cover of DuckDB in Action
liveBookmanning.com
Read DuckDB in Action on liveBook Get the book at manning.com

Related concepts

This explainer was generated by Manning's knowledge graph from the text of DuckDB in Action. The section links point to the parts of the book it draws on.