Clean your purchase data before your JD Edwards migration
Oracle's support policy for JD Edwards EnterpriseOne 9.2 has also been extended, following the same pattern as E-Business Suite, so most JD Edwards customers are moving on their own timeline rather than against a forced deadline.
Where the data breaks
- Category codes have been repurposed for many different things over the years, so the same code field means something different in different modules.
- One address book holds suppliers, customers and employees together, which makes it hard to isolate supplier data cleanly.
- Item descriptions are written differently at each branch, since JD Edwards never enforced a shared format across sites.
What it costs to leave it
Bad data doesn't stay behind when you migrate. It gets carried into the new system, where it causes extra test runs, a delayed go-live, and spend reports nobody trusts once the dust settles.
Where this fits in your migration
Plan and choose a partner
The business case is agreed and an implementation partner is chosen.
Assess the data
What exists, what to move, and what state it is in. Pearstop starts here, by checking and classifying your purchase and supplier data.
Design the new system
Processes, fields and how old data maps to new. Pearstop maps the cleaned data to the new system's fields.
Test the move
Before go-live, your partner tests the move by loading the data into a trial version of the new system, often several times. We hand over the cleaned data in the format the new system imports, so those tests don't fail on bad records.
Go live
The new system goes live for your teams.
Run
New invoices stay classified, so reporting doesn't slip back.
A mid-sized migration often takes six to twelve months; larger SAP and Oracle programmes often run a year or more. Data problems usually show up during the test runs, when there is least time to fix them.
Your teams, who own the data.
Keeps systems, access and integrations running through the move.
Sets up the new system.
Cleans and classifies the data.
The data that matters in a Oracle Fusion migration
- Address book records for suppliers, which need separating cleanly from customers and employees before they migrate.
- Category codes, which usually need replacing with one classification standard rather than carried forward as repurposed fields.
- The item master, where descriptions typically vary the most between branches.
What good data looks like
- One record per supplier, with no duplicates across entities.
- Every spend line labelled to one standard (UNSPSC or your own).
- Item and material descriptions written the same way.
- Units of measure consistent.
- Inactive suppliers and items flagged, not migrated.
- Every field mapped to the new system and documented.
- Suppliers separated cleanly from other address book records.
- Category codes replaced by one standard.
How Pearstop handles it
The data doesn't need to be clean first. AI classification handles the clear cases, and a person reviews anything uncertain. UNSPSC labels are mapped to Oracle Fusion's fields, in the format your migration needs. New invoices stay classified after go-live, so reporting doesn't slip back into the same mess.
Work with your partner
Most migrations run through an implementation partner. We work alongside them, or directly with you. Your partner sets up the new system; we give them clean, classified data in the format their import tools need. Nobody's scope changes, and your partner's team can spend their time on the new system instead of fixing records.
Why start early
Bring us in when you assess the data, not after the first test run fails. You find duplicates, free-text lines and missing categories while there is still time to fix them, nothing has to be cleaned twice, and the test runs go faster because the data goes in cleanly. Some problems only show up once you look closely at how the data behaves, which is why it pays to work with someone who handles this data every day.
Frequently asked questions
How do we clean the JD Edwards address book before migrating?
Separate supplier records from customer and employee records first, then de-duplicate suppliers across branches, since JD Edwards' single address book design makes it easy for these to get mixed together.
What happens to JD Edwards category codes in Oracle Fusion?
Category codes don't map one-to-one to Fusion's structure, particularly where they've been repurposed over the years, so it's usually better to replace them with one classification standard rather than force a direct mapping.
Related reading
Send us 200 lines
We'll classify a sample of your purchase data and send it back so you can see what good data looks like before you commit to cleaning the rest.
