View Layer
Stay organized with collections
Save and categorize content based on your preferences.
The heavily nested nature of FHIR resources and the Parque-on-FHIR schema
requires complex SQL queries that can make them difficult to work with for
analytics use cases.
A common approach to address this is to flatten the data into a set of views
(virtual or materialized) which can then be queried using simpler SQL
statements.
Using FHIR Data Pipes, there are multiple approaches for generating flat tabular
views for use in analytics applications to match the deployment requirements:
SQL queries to generate virtual views (outside the pipeline)
FHIR ViewDefinition resources to generate materialized views (within the pipeline)
which can be outputted in any tabular format (with current support for Parquet
and DB tables)
For both of these approaches, a set of "predefined views" for common FHIR
resources are provided. These can be modified or extended.
Go to developer documentation
Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.
Last updated 2024-11-26 UTC.
[null,null,["Last updated 2024-11-26 UTC."],[[["\u003cp\u003eFHIR data's nested structure necessitates complex SQL queries, making analytics challenging.\u003c/p\u003e\n"],["\u003cp\u003eFlattening FHIR data into views simplifies querying for analytics.\u003c/p\u003e\n"],["\u003cp\u003eFHIR Data Pipes offers two approaches for generating flat tabular views: SQL queries for virtual views and FHIR ViewDefinition resources for materialized views.\u003c/p\u003e\n"],["\u003cp\u003ePredefined views for common FHIR resources are available and customizable.\u003c/p\u003e\n"],["\u003cp\u003eExplore detailed information and implementation guidance in the developer documentation.\u003c/p\u003e\n"]]],["FHIR data's complexity necessitates flattening for analytics. This is achieved by creating tabular views, either virtual or materialized. Virtual views are created using SQL queries, while materialized views are generated via FHIR ViewDefinition resources, outputting in formats like Parquet or DB tables. FHIR Data Pipes supports both methods, providing predefined views for common FHIR resources that can be customized, to facilitate analytics applications. Further details are available in the developer documentation.\n"],null,["# View Layer\n\nThe heavily nested nature of FHIR resources and the Parque-on-FHIR schema\nrequires complex SQL queries that can make them difficult to work with for\nanalytics use cases.\n\nA common approach to address this is to flatten the data into a set of views\n(virtual or materialized) which can then be queried using simpler SQL\nstatements.\n\nUsing FHIR Data Pipes, there are multiple approaches for generating flat tabular\nviews for use in analytics applications to match the deployment requirements:\n\n1. SQL queries to generate virtual views (outside the pipeline)\n\n2. [FHIR ViewDefinition](https://build.fhir.org/ig/FHIR/sql-on-fhir-v2/StructureDefinition-ViewDefinition.html) resources to generate materialized views (within the pipeline)\n which can be outputted in any tabular format (with current support for Parquet\n and DB tables)\n\nFor both of these approaches, a set of \"predefined views\" for common FHIR\nresources are provided. These can be modified or extended.\n\nGo to [developer documentation](https://google.github.io/fhir-data-pipes/)"]]
