ETL processing β dimensions
ZTZenoti Docs Team Updated April 10, 2026 2 min readGuide
Dimension processing extracts reference data from your Zenoti system and transforms it into standardized dimension tables in your Analytics β Amazon S3 data warehouse.
Before you begin
- Your organization must have an active Analytics β Amazon S3 subscription.
- The initial full load must be completed before incremental dimension processing begins.
- You need access to your organization's S3 bucket to verify dimension data delivery.
Steps
Understand dimension processing stages
Analytics β Amazon S3 processes dimensions through two paths:
- SQL Server Step β Creates two core dimensions (
Bi_DimUserandBi_DimCenter) in the Analytics database using business transformation logic. - Redshift Dim/Fact Load β Creates the remaining 95 dimensions directly in the Redshift data warehouse from source system data.
This two-path approach means that the user and center dimensions are available earliest in the pipeline, while all other dimensions are built during the Redshift loading stage.
Understand incremental dimension refresh
The system automatically processes changed dimension records during each data refresh cycle:
- Change detection β The system identifies modified records since the last refresh. Only changed records are processed.
- Surrogate key generation β Each dimension table receives a BIGINT IDENTITY surrogate key column (WID β Warehouse ID) as the primary key. Natural business keys (GUIDs) are preserved for cross-referencing.
- S3 delivery β Updated dimension data is exported as CSV files to your S3 bucket, organized by PODName and GroupID folders.
Verify dimension table structure
All dimension tables follow a consistent structure:
- Surrogate key (WID) β Auto-generated warehouse ID for optimal join performance.
- Natural key β Original GUID or ID from your Zenoti system.
- Business attributes β Descriptive fields like names, codes, and status indicators.
- Multi-tenant fields β
PodNameandgroupidfor data segregation in S3 Data Share mode.
Track dimension processing status
Monitor dimension processing through the data refresh cycle:
- The scheduler triggers incremental sync at configured intervals (every 1β2 hours).
- The system processes changed dimension records before fact table processing.
- CSV files are placed in temporary
<groupid_processing>folders during processing. - Upon completion, folders are renamed to permanent
<groupid>structure.
Tips
- The system maintains 97 dimension tables across business domains including guests, centers, employees, services, products, memberships, campaigns, and more.
- Dimension tables are replicated across all warehouse nodes for optimal join performance with fact tables.
- Always process dimension table updates before fact table updates to maintain referential integrity.
Was this article helpful?
Your feedback helps us improve our documentation.