Data Management

Guidelines for organizing, sharing, and managing data in Elementum

Maximize your data’s value with proven strategies for organization, sharing, and management. These guidelines cover table and view choices, identifiers, integrity, change monitoring, and loading patterns. For related options by use case, see Data structure options.

New to data integration? Start with the CloudLink setup guide to establish your data connection first.

Data Sharing Fundamentals

Tables vs Views: The Foundation Decision

Choose between tables and views based on whether you need writes, how you filter columns, and performance. For a concise matrix of structure types by scenario, see Data structure options below.

Use Tables When

Write Operations Required

Use a table when data is updated through automations or user input:

• User form submissions
• Automation-driven status updates
• Real-time data modifications
• Transactional processing

Performance is Critical

Use a table when queries run often and latency matters:

• Dashboard data sources
• Real-time reporting
• Frequently accessed reference data
• Performance-sensitive workflows

Use Views When

Data Subset Needed

Use a view when you need to share only certain columns or filtered rows:

• Security-sensitive data with column restrictions
• Department-specific data views
• Filtered datasets based on business rules
• Simplified data presentations

Read-Only Access Sufficient

Use a view when no writes are required:

• Reporting and analytics
• Reference data lookups
• Audit and compliance views
• Data exploration and discovery

Special Considerations

Data exchange tables

When you need both read and write behavior and the presentation benefits of a view, use a view for display and a separate exchange table for updates.

Example: A customer view for display plus a customer_updates table for changes.

To create automations based on record changes (updates, new records, etc.) or track business entities, create an Element with your data instead.

Unique Identifiers

Each row in your data must have a unique identifier. If one does not exist, create it using:

• UUID functions
• Concatenated fields
• Incremental number assignment

Always verify uniqueness by running a validation query before linking.

Table Integrity Guidelines

Critical Changes

These changes cause “No Rows” errors:

• Renaming tables
• Renaming shared columns

Changes Requiring Review

Consult with Elementum before:

• Deleting columns
• Repurposing columns
• Using Create/Replace table commands

Low-Impact Changes

• Adding new columns (requires Elementum admin to update layouts)

Data Structure Options

Use Case	Recommended Option	Example
Read-only access	Regular table/view	Data Mining
Write access needed	Hybrid table	Reclaim License Data Exchange
Performance critical	Materialized view/hybrid table	-
Multi-table queries	Dynamic table	-

For conceptual guidance on tables versus views, see Tables vs Views above.

Change Monitoring

Choose your monitoring approach based on frequency:

Known Frequency

• Use data mining for hourly, daily, weekly, or monthly updates
• Row limits per data mine apply; see Field updates

Real-time Needs

• Use record created/updated triggers
• Enable CHANGE_TRACKING for automated updates (configuration details in Change tracking configuration)

Time-based Updates

• Use time-based triggers for known update schedules
• Elementum detects changes based on resource scheduler or data mine schedule

Field Updates

Automation Options

• Use create triggers for non-null initial values
• Use data mines for batch updates (100,000 row limit per data mine)
• Enable change tracking for real-time element updates (see Change tracking configuration)

Calculated Fields

Use for related item data:

STRING_AGG_UNIQUE(HANDLE."Field", ', ')  -- For text
MAX(HANDLE."Field")                      -- For numbers

Change Tracking Configuration

Capabilities

• Tables: Select, Update, Insert, Delete (with permissions)
• Views: Select only

Implementation Notes

• Enables automation triggers from tables and Elements
• Supports table modifications for joins and data mines
• Works best when update frequency is known so scheduling stays efficient

Data Loading Strategies

Incremental Loading

• Works directly with CHANGE_TRACKING
• Efficient for regular updates

Flush and Fill

• Requires duplicate checking
• Use search actions before create/update operations

Update Behavior Matrix (Snowflake)

The following describes trigger behavior when using Snowflake with change tracking:

CHANGE_TRACKING	Elementum Triggers	Snowflake Triggers
Enabled before config	Immediate	Based on scheduler
Enabled after config	Immediate	Won’t fire
Disabled	Immediate	Won’t fire

Next Steps

Use these guides to connect your warehouse, align table design with how Elementum uses data, and operationalize monitoring and loads.

Connect Snowflake

Run the setup script and connect your Snowflake account to Elementum

CloudLink in Elementum

Finish CloudLink configuration after your warehouse is connected

Data mining

Schedule batch pulls, respect row limits, and align mines with your change strategy

CloudLinks

Understand how CloudLinks expose warehouse data to apps and automations

Tables

Work with table objects, layouts, and how data surfaces in the workspace

Table types

Choose table types and patterns that match read, write, and performance needs