HomeGuidesSDK ExamplesAnnouncementsCommunity
Guides

Editing Columns / Attributes

Suggest Edits

Table Column Selection Assistant

What it does: Provides a comprehensive view of all available columns in your underlying data table when editing dataset attributes.

How to access:

  1. Go to Data page → Select dataset → Columns tab
  2. Click "Add" or edit an existing column
  3. Look for the new right-side panel showing available table columns

Features:

  • Complete column list: See all columns available in the underlying data table
  • Status indicators:
    • ✓ Check mark: Column is already configured as an attribute
    • "Missing" indicator: Column exists in table but not yet added as attribute
  • One-click copying: Copy button next to each column name for easy use in SQL expressions

Workflow benefits:

  • No more guessing which columns are available
  • Easy identification of unconfigured columns
  • Streamlined process for building complex SQL expressions
  • Reduced errors from mistyped column names

Enhanced Column Configuration

Dimension Value Count:

  • New column shows the count of unique values for each dimension
  • Helps understand data distribution and completeness
  • Useful for identifying high-cardinality dimensions that might impact performance

Sample vs. Total Context:

  • Better visibility into how much data you're working with
  • Compare sample sizes to total available data
  • Make informed decisions about sample limits for performance optimization

Data Validation and Integrity

SQL Syntax Validation

Real-time validation for dataset columns:

How it works:

  1. When editing column SQL expressions, syntax is validated in real-time
  2. Invalid syntax prevents saving and shows specific error messages
  3. Validation focuses on SQL syntax correctness, not column existence

What it validates:

  • ✅ SQL syntax correctness (parentheses, operators, functions)
  • ✅ Proper SQL structure and formatting
  • ❌ Does NOT validate if referenced columns exist in your data

Example scenarios:

  • Valid: SUM(revenue) / COUNT(*) - Proper syntax even if columns don't exist
  • Invalid: SUM(revenue / COUNT(*) - Missing closing parenthesis
  • Invalid: SUM revenue - Missing parentheses for function

Benefits:

  • Catch syntax errors before they cause runtime issues
  • Particularly valuable for complex calculated fields
  • Immediate feedback during column creation and editing

JSON Configuration Validation

Enhanced validation for dataset metadata:

Where it applies:

  • Dataset Properties → Metadata section (used for hierarchies and custom configurations)
  • Column-level JSON configurations

Features:

  • Real-time JSON syntax validation
  • Visual indicators (red squiggles) on syntax errors
  • Line-specific error feedback
  • Full-screen editing mode for complex configurations

How to use:

  1. Navigate to dataset Properties tab
  2. Scroll to metadata section at bottom
  3. Edit JSON configuration
  4. Validation occurs in real-time as you type
  5. Use full-screen mode for complex hierarchies

Benefits:

  • Prevents invalid JSON from breaking dataset functionality
  • Immediate feedback on syntax errors
  • Professional editing experience with syntax highlighting

Best Practices

Before Making Dataset Changes

  1. Check dependencies: Always review the Skill Dataset Dependencies table
  2. Identify impact: Note which skills use the dataset and how (inherited vs. direct)
  3. Navigate and review: Click into affected skills to understand usage context
  4. Plan changes carefully: Consider creating new datasets for major structural changes

When Building New Columns

  1. Use the column assistant: Review available table columns before starting
  2. Copy column names: Use the copy button to avoid typos in SQL expressions
  3. Start simple: Build complex expressions incrementally, validating syntax at each step
  4. Check samples vs. totals: Understand your data distribution before setting sample limits

For Complex Configurations

  1. Validate as you go: Take advantage of real-time validation for SQL and JSON
  2. Use full-screen mode: For complex JSON hierarchies, use the full-screen editor
  3. Test incrementally: Save and test small changes before building complex expressions
  4. Document relationships: Use descriptive names that reflect dataset dependencies

Data Governance

  1. Regular audits: Periodically review dataset dependencies to identify unused datasets
  2. Consolidation opportunities: Look for datasets that could be merged or simplified
  3. Performance monitoring: Use dimension value counts to identify performance optimization opportunities
  4. Change management: Always check dependencies before deprecating or significantly modifying datasets

Updated 9 days ago