SUPPORTED DATABASES

DVT speaks to eighteen database engines, and one rule covers them all: every engine works as a federation source, a federation target, and a seed target, with full dbt compatibility on your default target. During federation, Sling moves the data and DuckDB does the joining — the federation engine page explains that machinery.

This page is the overview. Every engine in the table links to its own page — connection fields, profiles.yml and sources.yml examples, least-privilege grants, and known limitations — so once you know your engine is here, click through and set it up.

FULL ADAPTER SUPPORT (PUSHDOWN + EXTRACTION + TARGET)

✓ verified means tested live against a real instance in DVT's verification suite. Eleven of these engines have gone further: every ordered pair among them — all 110 engine-to-engine transitions, both directions — has been walked live under a 27-command battery (flip, parse, seed, double build, snapshots, idempotence, row counts and named values against canonical truth, docs, sources, metadata, base views, native execution, zero-residual retract), with a green gate on every hop. The 0.2.52 release notes tell that campaign's story.

PYTHON MODEL SOURCES

No native adapter, no problem — Python models run locally and land their results in your DAG like any other model. That is how DVT reaches these systems: Python models guide.

SLING-ONLY CONNECTORS (SOURCE + FILE TARGET)

No SQL engine behind them — these work as extraction sources for federated models, and as bucket-style targets (results land as files).

BUCKET STORAGES (SOURCE + TARGET)

Cloud buckets and the local filesystem work as extraction sources or materialization targets. Data moves via Sling, compute runs in DuckDB.

Full bucket storages guide — profiles.yml setup, source/target examples, incremental extraction, supported formats.

FILE FORMATS

These four formats are the whole file surface — the same list applies to bucket reads, bucket writes, and seeds, so a bucket source and a seed of the same file give the same columns. Anything else is refused by name. Table formats such as Iceberg or Delta are engine configurations, not file formats.