Best Practices

Configuration recommendations for attributes, including data type selection, transformation expressions, and performance optimization.

Attribute Configuration

1

Review Discovered Attributes

Always review discovered attributes after scanning:

• Verify data types are correct

• Check for unexpected polymorphic versions

• Understand the structure of your data

• Identify attributes that may need special handling

2

Adjust Data Types

Edit Snowflake data types if inference is incorrect:

• Set appropriate precision and scale for numeric types

• Specify datetime formats for accurate parsing

• Use correct VARCHAR lengths

• Consider storage and performance implications

3

Use Alias Names

Create readable alias names for better SQL readability:

• Follow your organization's naming conventions

• Consider downstream consumers when naming

• Make column names self-documenting

• Avoid overly long or cryptic names

4

Handle Null Values

Set appropriate null value expressions:

• Consider business logic for missing data

• Use COALESCE or CASE statements appropriately

• Document null handling decisions

• Ensure null handling is consistent across related attributes

5

Configure Transformations

Use transformation expressions for data cleaning:

• Apply business logic during materialization

• Document transformation logic in comments

• Test expressions before deploying

• Keep transformations maintainable

---

Virtual Attributes

1

Naming Conventions

Use clear, descriptive names:

• Follow consistent naming patterns

• Avoid spaces in attribute names

• Make names self-explanatory

• Consider how names appear in generated SQL

2

Transformation Expressions

Write clear, maintainable SQL expressions:

• Document complex logic in comments

• Test expressions before deploying

• Use standard SQL functions

• Avoid overly complex nested expressions

3

Semantic Model Integration

Use W_QUESTION_CATEGORY appropriately:

• Create metrics and filters as virtual attributes

• Link virtual attributes to parent arrays when needed

• Ensure expressions are compatible with Cortex Analyst

• Test semantic model queries

---

Polymorphic Versions

1

Review All Versions

Check all polymorphic versions for an attribute:

• Understand why multiple versions exist

• Decide which versions to include in code generation

• Document decisions about active/inactive versions

• Monitor for unexpected activations

2

Handle Polymorphism

Use transformation expressions to handle type variations:

• Consider data quality implications

• Document polymorphic behavior

• Use type checking in expressions if needed

• Ensure transformations handle all active versions

3

Version Management

Keep active only versions you need:

• Monitor version status dates for schema changes

• Review inactive versions periodically

• Understand that inactive versions are ready for activation

• Don't delete inactive versions unnecessarily

---

Performance Considerations

1

Include In Code Gen

Only include attributes you actually need:

• Exclude attributes that won't be used

• Reduces Dynamic Table size and refresh time

• Improves query performance

• Simplifies maintenance

2

Cluster By

Use ADD_TO_CLUSTER_BY for frequently queried attributes:

• Improves query performance on large tables

• Consider primary keys and foreign keys

• Balance clustering with refresh performance

• Monitor query patterns

3

Data Type Sizing

Use appropriate precision and scale:

• Avoid over-sizing numeric types

• Balance between flexibility and storage efficiency

• Consider future data requirements

• Optimize for common use cases