Import from CSV

Most major entities have an import page that walks a CSV file through upload, column mapping, validation, an optional linking step, and a final confirmation. Your column headers are matched without regard to case against the names Ceum recognizes, so a hand-built CSV usually maps cleanly with little manual fixing.

Where it's available

• Clients — go to Import clients.
• Projects — go to Import projects.
• Tasks — open the Tasks page and use its import action.
• Invoices — go to Import invoices.
• Transactions — go to Import transactions.
• Subscriptions — go to Import subscriptions.
• Time entries — go to Import time entries.

Documents don't have an import flow.

Wizard steps

1. File upload — pick a CSV (some entities also accept XLS, or even PDF; subscriptions accept all three). A Download sample action on every page builds a small example file using your current statuses, types, and tags, so you have a real template to start from.
2. Review & edit — Ceum reads your column headers, maps them to the right fields, and shows the parsed rows in an editable table. Bad cells are highlighted; you can fix typos, remove rows, or restore previously removed ones right there.
3. Linking (optional, only when relevant) — for example, importing projects can link them to existing clients by name. This step is skipped automatically when there's nothing to link.
4. Confirm — a summary of what will be created. Errors block confirmation; warnings let you continue.
5. Commit — Ceum creates (or updates) the records and reports success or any per-row failures.

Header matching

Column headers are matched without regard to case against the names Ceum accepts for each field. For example, the transaction date column is recognized as either Date or date, and the tag column as any of Category, category, Tags, or tags. The first recognized name for each field is the one used in error messages and in the sample file.

Validation

The review step surfaces:

• Field-level errors — a required field is missing, an email or date is badly formatted, or a value isn't one of the allowed options.
• Cross-column rules — for example, an end date must follow its start date, and an end time must follow its start time.
• Cross-row rules — duplicate-name checks within the file you're importing and against records you already have (for example, client names must be unique).

Errors stop you from advancing; warnings don't.

Duplicate handling

Most imports default to creating new records. For entities that require unique names, such as clients, a row that collides with an existing record by name produces a warning (or is blocked). A few imports can update an existing record when the file includes its ID; otherwise every row creates a new record.

Tips and edge cases

• Download the sample first. Sample files are built from your live workspace, so the statuses, types, and tags in the example actually exist.
• Tags split on commas. A single cell of vat, eu, q3 becomes three tags, and any tag that doesn't exist yet is created for you.
• Use ISO dates. Write dates as YYYY-MM-DD to be safe; locale-specific formats aren't reliably understood.
• The linking step is conditional. If there's nothing to link — say, every project row already names an existing client — the wizard jumps straight to confirmation.

On mobile

Not available on the mobile app — manage from the web app.