Rust · Python · Apache Arrow · Data quality
dataprof
Bounded-memory profiling for files, DataFrames, and Arrow streams.
A beta Rust crate and Python package for understanding unfamiliar tabular data: shape, types, statistics, quality signals, and comparable reports. The CLI belongs to historical releases.
Updated from public sources ·
Broader inputs, a smaller Python install
The current Python wheel supports local files, dictionaries, byte buffers, async HTTP profiling, and Arrow-compatible inputs. Database connectors remain a source-build feature. Reports can be saved, compared, rendered as Markdown, or summarized for an agent.
Correctness before a prettier score
Recent work covers Arrow C Stream inputs, complete traversal of chunked tables, Parquet sample coverage, stable numeric aggregation, and consistent empty-source behavior. Reports disclose when no quality assessment was possible instead of presenting missing evidence as a perfect score.