Exaloop Cookbook
These recipes combine Exaloop components into complete application workflows: reconciliation reports, SQL calculations, schema migration, telemetry validation, caching, and transactional event processing. Each includes reproducible inputs, an observable result, and deployment considerations.
Start with the all-SDK distribution, then run
exaloop doctor. Run each recipe in its own empty working directory: examples
create local fixtures and overwrite their named output files. No customer data,
credentials, or paid services are needed for the file-based recipes.
Choose a Recipe
| Task | What you combine | Starting point |
|---|---|---|
| Reconcile invoices and settlements | CSV, typed business rules, Parquet exception reports | Local files; Data SDK |
| Screen a loan portfolio in SQL | DuckDB, a native risk rule, Data SDK reporting | Embedded SQL; no server |
| Migrate a billing event schema | Avro aliases and defaults, bounded Parquet output | Versioned file contracts |
| Validate Protobuf telemetry | Generated messages, presence checks, quarantine reports | Message-to-analytics ingestion |
| Cache a PostgreSQL lookup in Redis | Parameterized SQL, native Redis, expiring keys | Two development services |
| Normalize orders in a Kafka transaction | Generated Avro, validation, atomic output and offsets | Transaction-capable Kafka cluster |
| Profile, tune, and package a batch worker | Profiling, PGO, held-out evaluation, runtime packaging | A complete downloadable project |
Suggested Paths
For a first project, start with reconciliation, then move the custom calculation into a DuckDB query. For data engineering, work through schema migration and telemetry validation before adding Kafka delivery semantics. For an application service, start with the cache-aside lookup and decide its consistency contract before adding concurrency. Finish any path with profiling and staging checks on a representative workload.
File recipes include complete native programs and self-generated inputs. The Protobuf and Kafka recipes also include their schemas and generation commands. Service recipes identify the development resources they write and provide an observable result; compiling them alone does not verify broker or database connectivity. Use the task guides for environment-specific security and deployment.
The recipes demonstrate integration patterns rather than predict performance. They place custom rules or numerical work alongside file formats, databases, and messaging systems, while leaving relational operations in SQL where appropriate. Use representative data to evaluate the cost of each execution boundary.
For API lookup, use the reference. For rollout, monitoring, and rollback, use the production guide.