Merge CSVs to Parquet
Combine multiple CSV extracts into one Parquet dataset on macOS — ideal when logs, exports, or shard files land separately but need to load as a single table. DataEngUtils merges locally with Polars so you never concatenate sensitive rows in a browser uploader. Align schemas, drop duplicate headers, and emit a compact Parquet file ready for DuckDB, Spark, or warehouse COPY commands. The tool is tuned for data engineers who receive weekly dumps from vendors and want a repeatable offline merge step before validation. Because merges often involve overlapping columns from different vendors, you can inspect conflicts on disk before anyone outside your team sees the combined file.
Try it free
Available in the desktop app
Merge CSVs to Parquet processes local files with Polars in Rust. Multi-gigabyte Parquet and CSV workflows require the native macOS app — nothing is uploaded to the cloud.
SELECT
order_id,
customer_id,
created_at
FROM analytics.fct_orders
WHERE created_at >='2026-01-01'
How it works
1. Select CSV inputs
Pick the files to merge. Order matters when timestamps or partitions are involved.
2. Harmonize schema
Resolve column mismatches and type conflicts before writing combined output.
3. Export merged Parquet
Write a single .parquet file sized for faster downstream scans.
Why not use an online tool?
Online converters 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 locally — same convenience, none of the exposure.