Parquet Schema Evolution
Compare two Parquet schemas and get a compatibility verdict before you ship a pipeline change. DataEngUtils diffs the schemas locally in your browser, listing added, removed and retyped columns and flagging whether the change is backward or forward compatible for readers. It is the check you want during code review when a model is about to change shape, when a producer ships a new field, or when an old consumer might break. Because the comparison runs client-side on an Apache Arrow engine compiled to WebAssembly, the schemas and the underlying files never leave your machine.
Open it in the app
Parquet Schema Evolution runs in the DataEngUtils browser app on a Rust engine (Apache Arrow) compiled to WebAssembly — part of the Pro workbench, entirely client-side, and nothing is uploaded.
How it works
1. Open Parquet Schema Evolution
Open the tool in the DataEngUtils app — part of the Pro workbench, fully local.
2. Add both Parquet files
Load the before and after files so the schemas can be diffed.
3. Read the verdict
See added, removed and retyped columns with a backward and forward compatibility summary.
Why not use an online tool?
Online inspectors tools require uploading your Parquet and CSV files to a third-party server. For production work, that is a security and compliance risk. DataEngUtils runs a client-side engine in your browser — same convenience, none of the exposure.