Cloud Portal ONE Desktop ONE DQ&C ONE MDM ONE RDM ONE Runtime Server
User Community Service Desk Downloads

Split Step

Splits the data flow into three streams based on a ONE expression condition, while also providing a copy of all records.

This step is available for standalone plans, embedded plans, and transformation catalog items. It is not available for transformation rules.

Overview

The Split step evaluates each record against a condition and routes it to separate outputs while also providing a complete copy of all records. This is useful when you need to process subsets of data differently while still maintaining access to the full dataset.

Use this step to:

  • Route records to different processing paths while keeping a master copy.

  • Create separate branches for valid and invalid records while logging all records.

  • Build complex workflows where some operations need all data and others need filtered subsets.

Configuration

Define the condition

  1. In the step configuration, enter your condition using a ONE expression in the Condition field.

  2. The expression must evaluate to a Boolean value (true or false).

  3. Connect downstream steps to the appropriate outputs.

Outputs

The Split step has three outputs:

out_true

Records where the condition evaluates to true.

out_false

Records where the condition evaluates to false.

out_all

All input records, regardless of the condition result.

Records appear in both a conditional output (out_true or out_false) and in out_all. Be mindful of this when designing your data flow to avoid processing records multiple times unintentionally.

Condition examples

Split by validation status
email is not null and customer_id is not null

  • out_true: Valid records for normal processing.

  • out_false: Invalid records for error handling.

  • out_all: All records for audit logging.

Split by region
region = 'EMEA'

  • out_true: EMEA records for regional processing.

  • out_false: Non-EMEA records for different processing.

  • out_all: All records for global reporting.

Split by priority
priority = 'High' or amount > 10000

  • out_true: High-priority items for expedited handling.

  • out_false: Standard items for normal processing.

  • out_all: All items for complete inventory tracking.

Example: Audit logging with conditional processing

A common pattern uses Split to maintain an audit trail while processing data differently based on conditions.

Use case

Process valid orders normally, send invalid orders for review, and log all orders for compliance.

Configuration

  1. Condition: order_date is not null and customer_id is not null and total > 0

  2. Connect out_true → Normal order processing → Database Output

  3. Connect out_false → Add error flags → Error table

  4. Connect out_all → Audit logging → Audit table

Split vs Condition vs Filter

Aspect Split Condition Filter

Number of outputs

3 (out_true, out_false, out_all)

2 (out_true, out_false)

1 (matching records only)

All records output

Yes (via out_all)

No

No

Non-matching records

Routed to out_false

Routed to out_false

Discarded

Use case

Route with full data access

Route to different processing

Remove unwanted data

When to use Split vs Condition

Use Split when:

  • You need access to all records for logging, auditing, or reporting.

  • Different branches need the full dataset alongside filtered subsets.

  • You’re building compliance workflows that require complete data trails.

Use Condition when:

  • You only need to separate records into two groups.

  • You don’t need a copy of all records.

  • Simpler routing logic is sufficient.

Best practices

Be mindful of data duplication

Records flow to both conditional outputs and out_all. Design your plan to avoid unintended duplicate processing.

Use for audit scenarios

Split is ideal when you need to maintain a complete record of all data while also processing subsets differently.

Connect all outputs

Even if you don’t need all three outputs, consider where each should route to avoid orphaned data paths.

Document the purpose

Clearly describe why you’re using Split instead of Condition, especially regarding the use of out_all.

See also

Was this page useful?