Skip to content

Import files

Importing files is stage 1 of the RebelCore™ data flow:

Import (raw) → Create dataset (silver) → Tree (gold, vectorized) → Agent

This is where your raw source data enters the system. It’s quick — you create an import batch and drop files in. There’s no transformation yet; the structure work happens in the next stage, Create your dataset.

What you can import

RebelCore™ currently accepts:

  • .csv
  • .xlsx (Excel)

That’s it. Other formats (JSON, Parquet, databases, etc.) aren’t supported in the import flow yet.

Steps

1. Open the Imports page

From an open project, click Imports in the side menu.

2. Start a new import batch

  1. Click New import batch.
  2. Give it a name you’ll recognise later (e.g., Q4 2026 — initial data drop).
  3. Click Create.

A batch is just a container — it groups related files so you can curate them together.

3. Upload your files

Inside the new batch, click Upload files then either:

  • Drag-and-drop your CSV / Excel files into the upload area, or
  • Click Browse and pick them.

Each file shows its own progress bar. You can come back later and add more files to the same batch — nothing is final yet.

4. (Optional) Adjust import settings

Click Settings on the batch to tweak how RebelCore™ should parse the files (e.g., header row, encoding, separator). Defaults work for most well-formed CSV / Excel files.

5. Run the import

When all files are uploaded, click Run import. The batch moves to Processing while RebelCore™ parses the files. Wait until the status flips to Created.

What’s next

Importing only puts the raw files into RebelCore™. To turn them into something usable, continue to Create your dataset — the silver stage where you give the data structure.

If it fails

A failed batch shows a red status with a short error message. Common causes:

  • File isn’t .csv or .xlsx.
  • Inconsistent column structure across sheets/files that should match.
  • Network interruption during upload — re-upload the affected file.

For anything else, click into the batch to see the full error and reach out to support if needed.