LinuxCommandLibrary
GitHubF-DroidGoogle Play Store

pqrs

Inspect Apache Parquet files from the command line

TLDR

Print schema
$ pqrs schema [file.parquet]
copy
Print contents
$ pqrs cat [file.parquet]
copy
JSON lines output
$ pqrs cat [file.parquet] --json
copy
CSV output
$ pqrs cat [file.parquet] --csv
copy
First N records
$ pqrs head [file.parquet] --records [10] --json
copy
Row count
$ pqrs rowcount [file.parquet]
copy
File size breakdown
$ pqrs size [file.parquet]
copy
Random sample
$ pqrs sample [file.parquet]
copy

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
copy

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

RESOURCES

Copied to clipboard
Kai