Differences Between Batch and Transaction Processing of FHIR Resources in AWS HealthLake
Bundling FHIR resources - AWS HealthLake AWS Documentation To bundle
AWS allows users to bundle multiple FHIR resources into a single entity within AWS HealthLake.
Summary
AWS HealthLake offers two types of bundle processing for handling FHIR resources: 'Batch' and 'Transaction'. Users must select the appropriate method based on data characteristics and integrity requirements.
Key Players
Details
This article provides a technical guide on handling FHIR (Fast Healthcare Interoperability Resources) resources within AWS HealthLake. The core concept revolves around the 'Bundle' container structure used for data exchange. HealthLake supports two distinct processing models: 'Batch' and 'Transaction'. Batch bundles process each resource operation independently, meaning that even if some operations fail, successfully processed data remains (suitable when partial success is acceptable). Conversely, transaction bundles treat all resource operations as a single atomic unit; either all succeed, or the entire set fails and rolls back (essential for guaranteed referential integrity). This distinction guides usage: Batch mode is suitable for bulk uploads of unrelated data or historical records with potential validation issues. Transaction mode is recommended for scenarios requiring high referential integrity, such as creating or updating patient encounters where all associated data must be recorded simultaneously. Although both bundle types use the same underlying structure, their processing behavior is determined by the value assigned to the `type` field.
Technology Note
FHIR（Fast Healthcare Interoperability Resources）は医療データ交換の国際標準。このエントリの関連技術: FHIR
Original content copyright by respective publishers