Graph your tables. Index a billion GPS points. Travel back to version 42. Parse a FHIR document in the same query that joins an Iceberg snapshot. Commit the pipeline to Git and get lineage automatically. One SQL engine. Your cloud or your datacenter.
If you know PostgreSQL, you know Delta Forge. Same SQL semantics, same functions, same transaction model. Plus Delta Lake and Iceberg commands like VACUUM, OPTIMIZE, and time travel as first-class SQL -- not PySpark notebooks.
Full SQL environment with IntelliSense and live results
200+ functions, CTEs, window functions, PL/pgSQL procedures
Time travel, spatial joins, graph algorithms -- all in one query
A Spark cluster to touch Delta files. A graph database for network analysis. A geospatial service for location data. A lineage tool no one trusts. Format converters keeping everything talking. Delta Forge unifies all of it, runs it in SQL, and deploys on infrastructure you already own.
Delta Forge reads and writes both formats from the same SQL engine. No converters, no connectors, no ETL between ecosystems. Write to Delta, expose it as Iceberg automatically via UniForm. Time travel, ACID, schema evolution, deletion vectors -- all available as SQL commands.
SELECT region, SUM(revenue)
FROM orders VERSION AS OF 42
GROUP BY region;CREATE ICEBERG TABLE metrics (
day DATE, value DOUBLE
) LOCATION 's3://lake/metrics';15+ graph algorithms run as SQL table functions against your Delta tables directly. H3 geospatial indexing handles billion-point spatial joins in the same query. Combine PageRank results with a spatial filter and JOIN back to your lake data, all in one SQL statement. No graph database. No geo microservice.
SELECT c.name, pr.score
FROM cypher('network', $$
CALL algo.pageRank({dampingFactor: 0.85})
YIELD node_id, score RETURN node_id, score
$$) AS (node_id BIGINT, score DOUBLE) pr
JOIN customers c ON pr.node_id = c.id
ORDER BY pr.score DESC LIMIT 10;SELECT h3_cell_to_string(
h3_latlng_to_cell(lat, lng, 9)
) AS hex_cell,
COUNT(*) AS visit_count
FROM gps_points
GROUP BY hex_cell
ORDER BY visit_count DESC;Write a SQL file. Commit it. Delta Forge scans your repo, schedules it, and builds the full data lineage graph automatically from your source code. No drag-and-drop DAG builders. No manual lineage annotation. Every table transformation is traced, every dependency is known, and your pipelines are code-reviewed like everything else.
PIPELINE 'daily_revenue'
SCHEDULE '0 2 * * *'
NOTIFY ON FAILURE 'ops@acme.com'
RETRIES 3;-- Delta Forge reads your .sql files
-- and traces: raw_events
-- -> enriched_events
-- -> revenue_summary
-- Zero config. Zero annotation.Native Delta Lake engine with ACID transactions, time travel, and schema evolution.
Learn more →First-class Iceberg V1/V2/V3 support. UniForm interoperability. True format freedom.
Learn more →Common file formats, relational databases (PostgreSQL, MySQL, SQL Server, Oracle), and cloud services. Predicate pushdown, schema inference, parallel ingestion.
Learn more →15+ graph algorithms as SQL table functions. PageRank, community detection, shortest path. No separate graph database required.
Learn more →GDPR-compliant data protection. Hash, encrypt, mask, redact. All reversible via SQL.
Learn more →Native FHIR R4/R5, HL7v2, EDI/X12 parsing. Query clinical and supply chain data with SQL.
Learn more →Uber H3 hexagonal indexing. Location analytics, spatial aggregation, hierarchical resolution.
Learn more →200+ functions, PL/pgSQL, window functions, CTEs, and recursive queries.
Learn more →Desktop GUI, VS Code extension, CLI. IntelliSense, query visualization, live results.
Learn more →Query any previous version of your table by version number or timestamp. Audit trails and reproducibility built in.
SELECT * FROM customers VERSION AS OF 3
Add, rename, or reorder columns without rewriting data. Automatic schema enforcement on write.
Track row-level changes for incremental processing. Before and after images for every modification.
Efficient row-level deletes without file rewrites. Reduces write amplification and conflict potential.
Native V1/V2/V3 Iceberg support. Same SQL engine for both Delta and Iceberg tables. Partition evolution, hidden partitioning, and snapshot isolation.
Expose Delta tables as Iceberg via UniForm. One write path, both ecosystems. No data duplication, no ETL pipeline between formats.
200+ functions. Recursive CTEs. Window functions with ROWS, RANGE, and GROUPS frames. Lateral joins. PIVOT/UNPIVOT. Full subquery support.
SQL Reference →Full procedural language support. Variables, loops, cursors, exceptions, and RETURN NEXT for set-returning functions.
Procedure Guide →PostgreSQL, MySQL, SQL Server. S3, Azure, GCS. CSV, JSON, XML, Parquet, ORC, Avro. Predicate pushdown where supported.
Connector List →Scalar, aggregate, window, and table functions. Write in Python or JavaScript with Arrow integration.
UDF Guide →Role-based access control. Column-level grants. Dynamic data masking. Row-level security. Audit logging.
Security Docs →GDPR Article 25 compliant data protection. keyed_hash, encrypt, redact, mask, generalize transforms. Runtime or permanent.
See Demo →Desktop application built with Tauri and React. VS Code extension with query editing and results viewing.
Tools Guide →Open the GUI, pick a demo, run the SQL. Built-in assertions verify every result automatically.
Delta Lake, Iceberg V3, graph analytics, geospatial, and auto lineage -- unified in one SQL engine. Use the Desktop GUI, VS Code extension, or CLI. Runs on AWS, Azure, GCS, on-premises, or air-gapped.