The delta-forge-demos repository contains use cases across Delta Lake, Iceberg, graph, geospatial, healthcare formats, EDI, and common file formats. Each use case creates its own tables, runs asserted queries, and cleans up after itself. A passing run means the assertions held; a failing run returns the query, the expected value, and the actual value.
Launch the DeltaForge GUI and browse use cases by category
One click creates tables and loads the seed data
Assertions validate results against expected values computed before the test ran
Teardown drops everything so you can move to the next use case
ACID transactions, time travel, schema evolution, and table maintenance on your own storage
CRUD operations, MERGE patterns (SCD2, dedup, soft delete), time travel, change data feeds, partitioning, VACUUM, OPTIMIZE, Z-ORDER, deletion vectors, schema evolution, column mapping, and GDPR erasure patterns.
V1, V2, and V3 table specs with schema evolution, partition transforms, bloom filters, CRUD operations, snapshot isolation, and UniForm interoperability between Iceberg and Delta Lake.
Graph traversal, geospatial indexing, and advanced SQL patterns
Cypher pattern matching, PageRank, community detection (Louvain), betweenness and closeness centrality, shortest paths, BFS, DFS, triangle counting, and KNN similarity. Tested on real-world public datasets.
Uber H3 hexagonal spatial indexing, WKT polygon operations, point-in-polygon queries, fleet tracking, delivery routing, and coverage analysis across multiple resolution levels.
Parse healthcare, supply chain, and logistics standards directly in SQL
FHIR R4/R5 clinical resources, HL7 v2 patient and lab workflows, pseudonymisation, and PII lifecycle management patterns.
HIPAA claims (837/835), eligibility (270/271), claim status (276/277), X12 purchase orders, EDIFACT international trade, TRADACOMS UK retail, and EANCOM supply chain messaging. Several of these include independent Python proof scripts that verify results without using DeltaForge.
Read and query structured, semi-structured, and binary formats
Columnar analytics with recursive directory scanning, predicate pushdown, file-level filtering, mixed compression codecs, schema evolution, and row group statistics.
Deep path extraction, namespace handling, repeating element strategies, schema evolution across files, subtree capture for audit trails, and hierarchical data flattening.
Binary format deserialization with logical types, nullable unions, nested message flattening, repeated fields, enum decoding, and mixed compression.
Delimiter options, quoting modes, header detection, sheet and range selection, data cleansing, multi-file joins, and cross-table analytics.
Every use case is designed so you do not have to trust the output.
Every query includes ASSERT statements with pre-calculated expected values. Row counts, specific cell values, and aggregates are verified automatically.
Each use case runs independently. Setup creates tables, queries exercise features, teardown drops everything.
All seed data uses fixed values. Results are reproducible across platforms and runs. Several EDI use cases include independent Python proof scripts that verify expected values without using DeltaForge at all.
Download the GUI, open the gallery, and see the assertions pass on your install.