TechHacker Newsabout 2 hours ago
DuckDB Parquet Paging Performance Analysis
1 min read
DuckDB's file_row_number option pages through Parquet files 2.53x faster than OFFSET by skipping row groups without decompression. The speed advantage depends on having many row groups. OFFSET internally uses a similar rewrite for pages under 1,000,000 rows.
Level
Hype check
Tap to vote and see what everyone thinks.
