Skip to main content

Create a Splitter configuration

A Splitter Configuration defines how Records from one dataset should be routed into multiple outputs.

A good Configuration turns a business routing decision into repeatable logic that can be tested and reused.

Before you begin​

Before creating the Configuration, write down:

  1. The input dataset - Which File will be split?
  2. The output groups - What separate results should be created?
  3. The routing Fields - Which Fields determine where each Record belongs?
  4. The fallback plan - What should happen to Records that do not match?
  5. The review method - How will you confirm that each output is correct?

Build the Configuration in this order:

  1. define the output groups
  2. identify the Fields that drive routing
  3. write the Rule for each output group
  4. decide how to handle unmatched Records
  5. test with known examples for every group
  6. review counts and sample Records
  7. save with a clear name

Step 1: Define the output groups​

Start by naming the outputs you expect.

Examples:

  • East Region, Central Region, West Region
  • Approved, Rejected, Needs Review
  • Current Month, Prior Month, Future Dated
  • Standard Records, Exception Records

Clear output names make the Configuration easier to understand and the results easier to review.

Step 2: Choose routing Fields​

Identify the Fields that determine where each Record belongs.

Common routing Fields include:

  • region
  • department
  • owner
  • status
  • date
  • category
  • priority

Use Fields that are reliable enough to support the routing decision.

Step 3: Build Rules for each output​

Translate each output group into plain-language logic.

Examples:

  • Region equals East.
  • Status equals Approved.
  • Due date is before today and status is still Open.
  • Category is blank or not in the approved category list.

Keep the first version focused. Add edge cases only after the main routing logic is proven.

Step 4: Decide how to handle unmatched Records​

Do not ignore Records that fail to match expected groups.

Depending on the workflow, unmatched Records may need to:

  • go into an exception output
  • remain in an uncategorized review File
  • be corrected in the source data
  • trigger a Configuration update if the business Rule changed

Step 5: Test every output​

Use a sample that includes:

  • at least one Record for each expected output
  • Records near Rule boundaries
  • blank or unusual values
  • Records that should go to the fallback or exception output

Testing only the happy path can hide routing problems.

Naming guidance​

Name the configuration by business purpose.

Better examples:

  • Split Monthly Cases by Region
  • Route Vendor Records by Approval Status
  • Separate Invoice Exceptions from Standard Records

Weaker examples:

  • Splitter 1
  • New Split
  • File Routing

Save checklist​

Before saving, confirm:

  • each output group has a clear purpose
  • routing Rules match the business intent
  • unmatched Records have a defined destination or review path
  • every output has been tested
  • output names are understandable

Next step​

After the Configuration is ready, continue to Run Splitter.