the.com/parquet
The file format that made data scientists stop hoarding CSVs and start hoarding columns.
means A columnar storage file format that organizes data by column rather than by row, making it fast to read and compress for analytics workloads.
from Borrowed wholesale from the French word for a wood-block flooring pattern — 'parquet' originally named those geometric inlaid-wood floors made of small fitted pieces. The tech world reached for it because the file format arranges data in neat, tightly fitted blocks (columns of values packed together) rather than sprawling rows, much like parquet tiles fit together into an efficient, patterned surface. The French word itself traces to 'parc,' a small enclosed space — so underneath the data-engineering jargon is, fittingly, a tiny fenced-off plot.
- origin
- named after wooden floor patterns, also laid in tidy blocks
- storage style
- columnar, not row by row like CSV
- why it wins
- reads only the columns you need, skips the rest
- built by
- Twitter and Cloudera engineers, open sourced in 2013