Data Ingestion

The Validate Step

After Import completes, the Validate step checks that every staged record is correctly mapped and that the resulting dimension intersections are legitimate. Validation is a gate — data cannot be loaded to the Cube until it passes. This guide covers both validation types, the order they run, and how to fix errors.

Two Types of Validation

Validation runs in two sequential phases. You must resolve errors from the first phase before the second phase can surface its own errors.
diagramValidation Order of Operations

Loading diagram...

Phase 1: Transformation Validation

Transformation Validation checks that every source field has been mapped to a valid Cube dimension member through the Transformation Rules. If any source value does not match a rule in its Rule Group, it is flagged as a Transformation Error.
What you see: A list of unmapped source values, organized by dimension. Each row shows the source value that has no matching rule.
How to fix it:
  1. Identify the unmapped source value and the dimension it belongs to
  2. Navigate to the Transformation Rules for that dimension's Rule Group
  3. Add a new rule (One-to-One, Mask, Composite, etc.) that maps the source value to the appropriate target member
  4. Click Retransform to re-apply all rules without re-importing the file
  5. Click Validate again

Phase 2: Intersection Validation

Intersection Validation checks that the complete set of mapped target values forms a valid intersection in the Cube. Even if every individual dimension is correctly mapped, the combination might not make sense — for example, a Customer dimension member mapped to a Salary Grade account.
What you see: Records where the full intersection is invalid. The system identifies this before data reaches the Cube.
How to fix it:
  1. Click on the bad intersection to drill down and see the source GL/ERP account
  2. Right-click and select View Transformation Rules
  3. Each dimension appears on the left with its Target Value column — investigate each one to find which dimension has an incorrect mapping
  4. Select the incorrect Target Value, right-click, and select Edit Rule
  5. Choose the appropriate Target Value and save
  6. Close the windows and click Validate again
⚠️Warning
Intersection errors only appear after all Transformation errors are resolved. If you fix transformation errors and re-validate, you may then see intersection errors that were previously hidden. Plan for potentially two rounds of fixes.

Validation Status

StatusColorMeaning
Not yet validatedBlueThe user has not clicked Validate
All validations passGreenNo transformation or intersection errors — ready to Load
Transformation errorsRedSource values are unmapped — fix rules and Retransform
Intersection errorsRedMapped values form invalid combinations — fix the specific rule

The Retransform Workflow

Retransform is one of the most important features of the Standard Workflow. After editing Transformation Rules, clicking Retransform re-applies all rules to the existing staged source data — without re-importing the file.
The Retransform workflow:
  1. Takes the original source records from Stage (still stored from the Import step)
  2. Re-applies every Transformation Rule in the assigned profile
  3. Generates new target values
  4. Automatically re-validates
This means you can iteratively fix rules and validate until everything passes, all without touching the source file or re-uploading data.
ℹ️Info
Retransform is only available in the Standard Workflow. The Direct Load Workflow does not store source records in Stage, so there is nothing to retransform. If you need to fix mappings with a Direct Load, you must re-import the file entirely.

Standard vs. Direct Load Validation Behavior

BehaviorStandard WorkflowDirect Load
Transformation errorsPageable list — inspect all errorsLimited to 1,000 records per load — may need multiple imports to see all errors
Intersection errorsFull error analysisSame 1,000-record limit
RetransformAvailable — fix rules without re-importingNot available — must re-import
Error storagePersisted in Stage tablesStored in memory only (up to 50,000 total records)

Fixing Errors Efficiently

Tips for resolving validation errors quickly:
  1. Start with Transformation errors — These are always Phase 1. Get all source values mapped before worrying about intersections.
  2. Use View Transformation Rules (right-click) to see exactly which rules applied to a given record. This shows you the full mapping context.
  3. Use wildcard rules cautiously — A broad Mask rule like * can unintentionally capture values meant for more specific rules. Check the Order field on Composite rules.
  4. Check the processing log if error counts seem wrong — the log shows total rows parsed, rows transformed, and rows with errors.
  5. For recurring files, fix rules once and future imports will map correctly without intervention.