pqrs
Inspect Apache Parquet files from the command line
TLDR
Print schema
$ pqrs schema [file.parquet]
Print contents$ pqrs cat [file.parquet]
JSON lines output$ pqrs cat [file.parquet] --json
CSV output$ pqrs cat [file.parquet] --csv
First N records$ pqrs head [file.parquet] --records [10] --json
Row count$ pqrs rowcount [file.parquet]
File size breakdown$ pqrs size [file.parquet]
Random sample$ pqrs sample [file.parquet]
SYNOPSIS
pqrs [*flags*] *subcommand* [*args*]
DESCRIPTION
pqrs is a Rust command-line utility for inspecting Apache Parquet files (roughly "parquet-tools in Rust"). It is built on the Arrow/Parquet Rust crates and replaces much of the deprecated Java parquet-tools workflow for schema, content, and size inspection.Install from GitHub releases, cargo install pqrs, or Homebrew (manojkarthick/tap/pqrs).
PARAMETERS
cat *files*...
Print records. --json, --csv, --csv-data-only / --no-header control format. Recurses into directories.head *file*
First N records (--records). Supports --json.schema *files*...
Print Parquet schema.rowcount *files*...
Print row counts.size *files*...
Print size information.sample *file*
Random sample of records.merge *files*...
Merge files by concatenating row groups into another Parquet file.-d, --debug
Debug output.
INSTALL
nix profile install nixpkgs#pqrs
CAVEATS
CSV output is not supported for files with Struct or Byte fields. Nested data is better viewed as default or --json. Very large files may need sampling rather than full cat.
SEE ALSO
parquet-tools(1), duckdb(1), arrow-tools(1)
