Excel to Parquet converter

Convert an Excel workbook to Apache Parquet in your browser. The first sheet is used; types are inferred. Files never leave the page.

How the Excel to Parquet conversion works

Drop an .xlsx workbook. The first sheet is read locally — header row becomes column names, date cells are preserved as real datetimes — and after a preview you download a typed .parquet with Snappy, GZIP, or no compression. Spreadsheets become something pandas, DuckDB, Spark, and warehouses ingest natively.

How cells become typed columns

• TRUE/FALSE → BOOLEAN
• whole numbers within ±2.1B → INT32, larger → INT64
• decimals → DOUBLE
• date/time cells → TIMESTAMP (real dates, not serial numbers)
• text or mixed → STRING
• Formulas convert as their last computed value. Empty and merged-away cells become nulls.

Limits

Only the first sheet converts; the header row is required. Very large workbooks parse in memory — for bulk work: pandas read_excel().to_parquet().