I see people using CSV exports, n8n or Zapier flows, and API scripts for this, and everyone seems to do it a bit differently. What’s your setup, and what’s the most annoying part of it?
I’d like to hear real workflows from real-world experience first before deciding on anything.
Thanks community
Not a long-running setup on my side, but here’s what tends to matter once you go past a one-time CSV import:
Decide which side is the source of truth first. Two-way sync is where most of the pain comes from (edits on both sides, conflicts, loops). One-way, Sheet to NocoDB or the other way, is a lot calmer.
Give every row a stable ID column in the sheet. Then n8n (it has both a Google Sheets node and a NocoDB node) or a small API script can do “find by ID, update if it exists, create if not” instead of creating duplicates every run.
Deletes are the annoying part. Most flows only catch new or changed rows, so a row deleted in the sheet just stays in NocoDB. A soft “archived” checkbox column is easier than real deletes.
Watch column types. Dates, currency and checkboxes come out of Sheets/Excel as text more often than you’d expect, so map them explicitly.
Sheet triggers in n8n/Zapier poll, so expect a short delay rather than instant updates.
If Excel is a one-off upload, NocoDB’s built-in import is honestly enough. If people keep editing the sheet daily, that’s when the ID + upsert flow pays off.