← All Releases

v0.2.3

2026-06-12

SEEDS, SUPERCHARGED

dvt seed grows into the simplest bulk loader you'll use: every Sling format, any target including buckets, and column names that behave identically on every engine.

ADDED

  • ▸Seed formats beyond CSV: tsv, parquet, json, jsonl, avro, arrow, xlsx — drop the file in seeds/ and run.
  • ▸dvt seed --target loads seeds to any profiles.yml output, including cloud buckets (as parquet objects).
  • ▸Python models: dbt-style .py models executed locally — REST APIs, MongoDB, ML inference — with results landing on any target through the same load pipeline.

CHANGED

  • ▸Column names are always snake_cased on seed loads ('Transaction Date' → transaction_date) — the same seed now loads identically on PostgreSQL, Oracle, Databricks, and everything in between.

Get it: pip install dvt-core==0.2.3 · wheels on PyPI for macOS (arm64 + x86_64) and Linux, Python 3.10–3.12.

← v0.2.2v0.2.5 →