Query Engine
Interactive SQL across any source at warehouse speed.
- Trino-backed federated queries
- Sub-second cache hits
- ANSI SQL with Iceberg time-travel
The platform
One platform. Every data workload.
DXData is a single lakehouse for interactive SQL, pipelines, governance, and observability. Built on Trino, Iceberg, and Nessie so your data stays open, portable, and yours.
// six capabilities, one surface
Everything you need, expanded.
Six first-class capabilities share the same catalog, the same permissions, and the same audit trail. No glue code, no integration tax.
Data Catalog
Every table, column, and lineage edge documented automatically.
- Auto-inferred schemas and lineage
- Column-level search and tagging
- Git-style change history
Pipeline Orchestration
Transformations that build themselves from visual graphs.
- Declarative YAML or visual DAG
- Incremental materialization with backfills
- Retries and SLA alerts built-in
Governance & RBAC
Row, column, and table-level controls with full audit trails.
- Attribute-based access policies
- SOC 2, HIPAA, and PCI DSS templates
- Immutable audit log for every query
Observability
Know exactly what is slow, what is stale, and what is broken.
- Query-level tracing with flamegraphs
- Dataset freshness and quality metrics
- OpenTelemetry export
Git-like Branching
Branch your data like your code — zero-copy and reversible.
- Named branches with write-audit-publish
- Tag-based rollbacks
- PR-style data review workflows
query_workspace.sql
-- federated across Postgres, S3 Iceberg, and Kafka
SELECT o.region,
SUM(o.total) AS revenue
FROM iceberg.sales.orders o
JOIN postgres.crm.customers c USING (customer_id)
WHERE o.created_at > '2026-01-01'
GROUP BY o.region;
4 rows · 218ms · 3 sources federated
cache hit// query.engine
One SQL surface over every source you own.
DXData runs a managed Trino cluster in front of your lakehouse, operational databases, and streaming systems. Queries plan across sources transparently, with adaptive join ordering and columnar pushdown where the engine supports it.
The result: dashboards render in under a second, analysts stop waiting on overnight ingestion, and nothing has to be copied to a separate warehouse to be useful.
- Iceberg time-travel, hidden partitioning, and schema evolution — natively
- Smart caching with per-dataset TTL and invalidation
- Cost-based optimizer with workload isolation per team
pipelines/orders_daily.yaml
Ingest
orders_rawClean
dedupe + castEnrich
join customersAggregate
daily_revenuePublish
BI martschedule: every 1hlast run: 4m ago · 0 failures
// pipelines.native
Pipelines your team can actually read.
Describe transformations as declarative YAML or sketch them in the visual DAG editor — both compile to the same execution plan. Incremental models track watermarks automatically, so backfills are a one-click operation instead of a weekend.
Runs are observable end-to-end: every step emits structured logs, lineage edges, and SLA signals that feed directly into your on-call.
- Backfill any window without duplicating rows
- Retries, circuit breakers, and failure budgets per stage
- Native integration with the catalog for lineage and docs
branch: experiment/customer_360
main
experiment
preview
// catalog.versioned
Branch your data like your code.
Powered by Nessie, every commit to the catalog is a pointer — not a copy. Spin up an isolated branch to test a schema change, run a backfill, or review a dataset before it touches production, then merge or discard it atomically.
Tags make rollbacks a first-class operation, and PR-style review flows let data consumers approve changes before they land on main.
- Zero-copy branches backed by Iceberg snapshots
- Write-audit-publish gates with policy hooks
- Time-travel queries across any branch or tag
// connectors
Plugs into everything you already run.
100+ native connectors for databases, warehouses, SaaS tools, object stores, and streams. Query in place or materialize into Iceberg — your call.
Databases
Postgres
MySQL
MongoDB
Oracle
SQL Server
Cassandra
Warehouses
Snowflake
BigQuery
Redshift
Databricks
Synapse
ClickHouse
SaaS
Salesforce
HubSpot
Stripe
Zendesk
Segment
Intercom
Files
S3
GCS
Azure Blob
Iceberg
Parquet
Delta Lake
Streams
Kafka
Kinesis
Pulsar
Pub/Sub
NATS
RabbitMQ
// architecture
A single layer between your apps and your data.
DXData sits in the middle — not as another silo, but as a thin, open layer that your tools already understand.
Apps & Consumers
BI Tools
Notebooks
Apps & APIs
ML / AI
DXData Platform
Query Engine
Catalog
Orchestration
Governance
Observability
Branching
Storage & Sources
Iceberg Tables
Object Storage
Operational DBs
Streams
// alternatives
How DXData stacks up.
vs Snowflake
Query any source without ingesting first — and keep your data on open Iceberg tables.
See full comparisonvs Databricks
Git-style branching and SQL-first workflows without a notebook-centric mental model.
See full comparisonvs BigQuery
Portable, open storage you own and flat pricing that never surprises finance.
See full comparisonvs Redshift
Federated queries across warehouses, lakes, and OLTP — no cluster resizing involved.
See full comparisonShip faster, own your data
Modern data platform, no vendor lock-in.
Everything in DXData runs on open standards you can walk away with — Iceberg tables, Nessie history, standard SQL.