All topics
34 / 60
Integrationintermediate

File-Based Data Import (FBDI)

Explain how FBDI enables bulk data loading into Fusion Cloud using standardized Excel templates

Fusion gives customers no database access, so EBS's insert-into-the-interface-table pattern is not available. FBDI is the replacement.

FBDI is like shipping goods through a standardized freight container system: you must pack your cargo (data) into the exact container dimensions and labeling format the shipping company (Oracle Fusion) requires, drop it at their loading dock (UCM), and then their own crane and forklift operators (ESS import process) take it from there — you can't just walk your goods directly onto the ship (database).

Key Concepts

1
The flow.
    Oracle's Excel template
        |  fill it in, run the macro
        v
    a .csv.zip in the expected structure
        |  upload
        v
    UCM (Fusion's content repository)
        |  submit the import scheduled process
        v
    interface tables -> application tables
2
Conceptually the same staging-then-import pattern as EBS. The difference is that the staging happens through a file and a repository rather than a table you can write to.
3
Use Oracle's template. There is one per business object — Journals, Suppliers, Items, AP Invoices — and the column order and names are fixed.
4
The macro builds the CSV and the zip. Hand-building a file that nearly matches the template produces a process that fails at row one with an unhelpful message, because the loader positions by column.
5
Automating it removes Excel entirely. For a recurring interface:
6
  1. Generate the CSV programmatically, matching the template's columns
  2. Zip it in the expected structure
  3. Upload to UCM via REST
  4. Submit the ESS job via REST
  5. Poll for completion and read the result
7
No manual step, which is what a nightly feed needs. Excel is for one-off conversions.
8
Errors come back as a report, not as an exception. The import completes, and some rows may have failed:
9
    processed:  10,000
    succeeded:   9,840
    failed:        160   -- with reasons, in the error report
10
That report is the thing to read. A conversion that "completed successfully" and is short by 160 records is a report nobody opened — the same failure mode as EBS interface tables, in a different place.
11
When to use which.
    REST API   tens to hundreds of records, immediate feedback
    FBDI       thousands upwards, a conversion or a scheduled feed
12
Loading fifty thousand journals through REST means fifty thousand calls against a rate limit. That is the distinction worth stating.
13
What the interviewer is probing.1. "What is the FBDI flow?" Probing: the mechanics. Stalls: "Upload a file." Moves up: fill the delivered spreadsheet template, generate the CSV and zip, upload to UCM, then run the load interface and import processes.
14
2. "Why use FBDI rather than REST?" Probing: the fit. Stalls: "It is older." Moves up: volume — it is the supported bulk path and uses the same validation as the delivered import programs.
15
3. "Where do you find out why rows failed?" Probing: the practical skill. Stalls: "The log." Moves up: the import process output and error report, which names the row and the validation that rejected it.
16
4. "What is the main project risk with FBDI?" Probing: experience. Stalls: "None." *Moves up:* templates change between quarterly updates, so a conversion built on last quarter's template can silently map columns wrongly.
Use cases:fbdiucmdata-conversionbulk-importess