Skip to Content
MappingsMappings

Mappings

Mappings are the backbone of Wave. Everything — from codegen to test cases — is rooted in your mapping logic.

Wave’s superpower is creating mappings from any data source — uploaded PDFs, API documentation, legacy database exports, vendor specifications, custom spreadsheets, or code repositories. While other tools need clean, standardized data, Wave thrives with messy real-world sources. This is where all the “how do I get from here to there?” logic lives.

How Mappings Work

  1. Wave analyzes your sources: Upload a PDF spec, paste API docs, or point Wave at any data source — it extracts field definitions automatically.
  2. AI generates intelligent mappings: Wave creates initial field connections with healthcare context and transformation logic built-in.
  3. Refine and review:
    • Use the mapping UI to tweak connections, add business rules, or resolve ambiguities
    • Or just ask Wave in chat to update mappings for you (“map gender, but transform values to FHIR codes”)
  4. Export or use in codegen:
    • Download mappings as .xlsx for validation or handoff
    • Generate ETL pipelines based directly on the mapping — never lose track of what’s happening under the hood

What Makes Wave’s Mappings Different?

  • Works with ANY data source — PDFs, APIs, legacy exports, vendor docs, custom formats
  • Mappings drive code generation — not just documentation, but actual implementation
  • Healthcare-aware intelligence — understands medical coding, FHIR extensions, concept mappings
  • Live collaboration — PMs, engineers, and SMEs can all view/edit the mapping, chat about it, and iterate together

Real-World Mapping Examples

Example 1: PDF Specification to API

  • Source: 47-page vendor PDF with custom patient fields (PatientUniqueId, DOB_MMDDYYYY)
  • Target: REST API with nested JSON (patientData.identifiers.primary)
  • Wave’s mapping: Extracts custom formats, handles date conversions, validates checksums

Example 2: Legacy Database to FHIR

  • Source: 15-year-old PostgreSQL schema with cryptic field names (pt_dob, gender_cd)
  • Target: FHIR R4 Patient resource with US Core extensions
  • Wave’s mapping: Reverse-engineers schema, maps to standard codes, handles value transformations

Example 3: Vendor Spreadsheet to Internal System

  • Source: Insurance company’s 200-column claims export with custom codes
  • Target: Your internal analytics database
  • Wave’s mapping: Parses complex headers, understands business logic, generates lookup tables

How Mapping Drives Code Generation

When you ask Wave to generate code (“create an ETL pipeline for this data”), it always consults your precise field mappings — so you get implementations that reflect your exact specifications, not AI guesswork.

  • Change the mapping, regenerate code, done
  • Every transformation is traceable back to your mapping decisions
  • Team stays synchronized — mapping is the single source of truth

Pro Tips

  • Already have mapping docs? Upload spreadsheets, data dictionaries, or documentation — Wave will parse and generate live mappings
  • Complex transformations? Wave handles value lookups, data type conversions, business rules, and edge cases
  • Version control? Mappings track changes and maintain audit trails for compliance

Next up: How to collaborate with Wave’s AI →

Data CatalogChats