Skip to content

Latest commit

 

History

History
38 lines (30 loc) · 1.81 KB

File metadata and controls

38 lines (30 loc) · 1.81 KB

ExcelBench Dashboard

Profile: xlsx | Generated: 2026-02-17T03:47:59.231174+00:00

Combined fidelity and performance view. Fidelity shows correctness; throughput shows speed. Use this to find the right library for your needs.

Library Comparison

Library Caps Modify Green Features Pass Rate Best For
openpyxl R+W Rewrite 16/16 100% Full-fidelity read + write
xlsxwriter W No 15/16 99% High-fidelity write-only workflows
rust_xlsxwriter W No 14/16 97% General use
wolfxl R+W Patch 14/16 97% General use
xlsxwriter-constmem W No 12/16 93% Large file writes with memory limits
xlwt W No 4/16 64% Legacy .xls file writes
openpyxl-readonly R No 3/16 23% Streaming reads when formatting isn't needed
pandas R+W Rebuild 3/16 20% Data analysis pipelines (accept NaN coercion)
pyexcel R+W Rebuild 3/16 21% Multi-format compatibility layer
tablib R+W Rebuild 3/16 21% Dataset export/import workflows
pylightxl R+W Rebuild 2/16 19% Lightweight value extraction
python-calamine R No 1/16 17% Fast bulk value reads
polars R No 0/16 15% High-performance DataFrames (values only)

Key Insights

  • Fidelity leaders: openpyxl (16/16 green features)
  • Abstraction cost: pandas wraps openpyxl but drops from 16 to 3 green features due to DataFrame coercion
  • Optimization cost: xlsxwriter constant_memory mode loses 3 green features for lower memory usage

Best Adapter by Workload Profile

Workload Size Best Read Adapter Best Write Adapter
small
medium
large