// the find
kaushiksrini/parqeye
Peek inside Parquet files right from your terminal
A terminal UI for inspecting Parquet files — schema, metadata, row group stats, and actual data rows, all navigable with keyboard shortcuts. Fills a real gap: Parquet is opaque binary and the usual alternatives are either heavyweight (DuckDB shell, Python/pandas) or don't show you the physical layout. Aimed at data engineers who spend time in the terminal.
Tab-based layout covering all four things you actually want to see in a Parquet file: data, schema, file metadata, and row group distribution. Built on Rust + ratatui, so it's fast to start and won't OOM on a 10 GB file the way a naive pandas read_parquet would. Homebrew and crates.io installs mean zero dependency friction. Actively maintained — last push yesterday, CI green.
No lazy/streaming yet (acknowledged in TODOs) — it loads the whole file into memory, which is a problem for the large files where this tool would be most useful. No column filtering or predicate pushdown in the visualize tab, so you can't easily find specific rows. Forks are low (12) relative to stars (668), which suggests people are using it but nobody is contributing — bus factor of one. No support for partitioned datasets or directories of Parquet files, which is how Spark/Hive output almost always lands.