Skip to main content
The CSV import on the Links page creates new Links and updates existing ones in one pass, up to 10,000 rows at a time. Every import runs as a preview first: QRtub reads the whole file, tells you how many rows would be created, updated or rejected and why, and writes nothing until you confirm. Use it when the addresses already exist somewhere else — a list of asset numbers from your maintenance system, a spreadsheet of destinations, an export you edited.

​
The columns

This is exactly the shape the Links Export produces, so the fastest way to get a valid file is to export, edit, and import the same file back. Every column is read as text, which matters for numbered slugs: leading zeros in cra0042 survive rather than being turned into 42. destination_url can be a URL template with field bindings, like app.com/inspect?id={{item.assetID}}. Bindings are inserted exactly as stored with no URL-encoding, so a field containing a space or an & produces a broken address — check the fields you reference before importing hundreds of rows that use them.

​
What each type needs on a create

Random — leave url blank. QRtub mints the slug. A random row with a url is only ever an update; if that url matches nothing, the row is rejected rather than guessed at, with a message saying so. Numbered — url is required, and it must match a numbering pattern your team already owns. cra0042 is accepted if you hold cra + 4 digits; otherwise the row is rejected with “does not match any numbering pattern owned by this team”. The import does not claim new patterns for you. Custom — url is required and must satisfy the custom-slug rules: 3–50 characters, lowercase letters, digits, hyphens and underscores, not a reserved word, not already in use, and not inside a numbered pattern your team owns.

​
How a row becomes an update instead of a create

By its url. If the slug already exists on one of your team’s Links, the row updates that Link instead of creating one — whatever link_type says. Matching follows each type’s own case rules: numbered and custom slugs match without regard to case, random slugs match exactly. Only your team’s Links are eligible. If you belong to several teams and a slug in the file exists in a different one, it is never treated as an update to that other team’s Link. On an update, columns absent from the file are left alone, so a two-column file of url,destination_url will not wipe everyone’s fallback settings. A column that is present but blank is treated as an instruction: a blank destination_url clears the destination, and a blank fallback_url or fallback_message clears that value.

​
The dry run

Choosing Import and picking a file runs the preview automatically. It performs the same classification and the same checks as the real thing — pattern matching, reserved-by-pattern, the random-slug rule, and duplicate detection — without writing anything, so the created and updated counts you see are honest rather than optimistic. Two rows creating the same new slug are caught here too: the first is counted, the second is rejected as a duplicate. Without that, the preview would promise two Links and the commit would deliver one plus an error. Confirm, and QRtub re-posts the same file to commit it. Rows that fail at this point — a custom slug that turned out to be taken, for instance — are reported individually and do not stop the rest of the import. You get a count of what landed and a per-row list of what didn’t, so you can fix those rows and re-import just them.

​
Limits

  • 10 MB per file, 10,000 rows, and the filename must end in .csv
  • Rows are processed one at a time, so a very large file takes a while — leave the tab open
  • The importer cannot assign Links to Items, add them to a print batch, or claim numbering patterns; it handles the Link’s own slug, destination, fallbacks and active state only