HCM Data Loader and HCM Spreadsheet Data Loader sit one letter apart and solve overlapping problems, which is exactly why beginners reach for the wrong one. Choose well and a load is routine. Choose badly and you'll spend days forcing a tool to do something it was never built for. The good news: the decision is simple once you know what each is actually for.
HCM Data Loader (HDL) — the bulk workhorse
HDL is file-based and built for scale. You construct .dat files — pipe-delimited text describing business objects — compress them, and load them through Data Exchange. It handles virtually every business object, very high volumes, and complex hierarchies, which makes it the tool for initial migrations and large mass updates. It's powerful and precise, but unforgiving: the file structure, the metadata and merge lines, the attribute names and the keys all have to be exactly right.
HCM Spreadsheet Data Loader (HSDL) — the friendly front door
HSDL is spreadsheet-based and sits on top of the same engine. You pick a template, download a spreadsheet, fill it in, and upload. That makes it ideal for functional and business users, smaller volumes, and ongoing incremental loads of common objects. The trade-off is reach: you're limited to the objects and templates exposed as spreadsheets, so it's less flexible than raw HDL.
HSDL runs on HDL underneath — same business objects, same validation. So the skills transfer directly: master HDL and HSDL becomes a friendlier door into the same room. Time spent understanding HDL is never wasted.
Choosing between them
Four quick questions usually settle it:
- Volume? Huge → HDL. Modest → HSDL.
- Who runs it? A technical team → HDL. A functional or business user → HSDL.
- Object coverage? An obscure or complex object → HDL. A common object with a template → either.
- What kind of load? Initial migration → HDL. Periodic top-ups → HSDL.
The usual culprits are data and sequencing, not HDL or HSDL: source keys versus GUIDs, loading objects out of dependency order (organisations before assignments), date-effective sequencing, and file encoding. Diagnose the data before blaming the loader.
Always run a small batch in a lower environment first, then read the load results carefully. Oracle tells you exactly which line failed and why — that error file is the fastest teacher you have.
Key takeaways
- HDL = bulk, file-based, every object, technical users, migrations.
- HSDL = spreadsheet-based, friendlier, smaller volumes, functional users, common objects.
- HSDL sits on top of HDL — same engine, narrower surface.
- Most load failures are data, ordering or date-effectivity issues, not the tool.
Know the two for what they are — one engine, two front doors — and data migration stops being a source of dread and becomes one of the most reliable skills you can offer a project.