vroom
Read and Write Rectangular Text Data Quickly
The goal of 'vroom' is to read and write data (like 'csv', 'tsv' and 'fwf') quickly. When reading it uses a quick initial indexing step, then reads the values lazily , so only the data you actually use needs to be read. The writer formats the data in parallel and writes to disk asynchronously from formatting.
README
# Benchmarks for vroom The benchmarks are run with a makefile, run `make` to run them. ## Personal notes on running benchmarks on AWS ### Starting up - Create new volume from previous vroom snapshot - `io1` volume type - 2500 IOPS - Attach volume to instance - `/dev/sda1` ### Attach to instance `ssh vroom-bench` When you first start there may be some unattended upgrades installing, use top to monitor and wait until they are done. ### running ``` make -j 1 \ TAXI_INPUTS='$(wildcard ~/data/trip_fare*csv)' \ FWF_INPUT=~/data/PUMS5_06.TXT \ BENCH_LONG_ROWS=1000000 \ BENCH_LOG_COLS=25 \ BENCH_WIDE_ROWS=100000 \ BENCH_WIDE_COLS=1000 ``` ### Tearing down - Shut down instance - Detach volume - Create snapshot - Delete volume
Versions across snapshots
| Version | Repository | File | Size |
|---|---|---|---|
1.7.1 |
latest linux/jammy R-4.5 | vroom_1.7.1.tar.gz |
974.1 KiB |
1.7.1 |
latest source/ R- | vroom_1.7.1.tar.gz |
602.2 KiB |
1.7.1 |
latest windows/windows R-4.5 | vroom_1.7.1.zip |
1.2 MiB |
1.7.1 |
2026-04-18 linux/jammy R-4.5 | vroom_1.7.1.tar.gz |
974.1 KiB |
1.7.1 |
2026-04-18 source/ R- | vroom_1.7.1.tar.gz |
602.2 KiB |
1.7.1 |
2026-04-18 windows/windows R-4.5 | vroom_1.7.1.zip |
1.2 MiB |
1.7.1 |
2026-04-12 linux/jammy R-4.5 | vroom_1.7.1.tar.gz |
974.1 KiB |
1.7.1 |
2026-04-09 linux/jammy R-4.2 | vroom_1.7.1.tar.gz |
964.6 KiB |
1.7.1 |
2026-04-09 linux/jammy R-4.5 | vroom_1.7.1.tar.gz |
974.1 KiB |
1.7.1 |
2026-04-09 windows/windows R-4.2 | vroom_1.7.1.zip |
1.2 MiB |
1.7.1 |
2026-04-09 windows/windows R-4.5 | vroom_1.7.1.zip |
1.2 MiB |
1.5.7 |
2022-02-12 windows/windows R-4.1 | vroom_1.5.7.zip |
1.9 MiB |