Sync DuckDB to Airtable on a Schedule

Automation · DuckDB to Airtable

Keep Airtable in step with your DuckDB file, on a schedule

Write the SQL once. Query Streams runs it on your schedule, anything from every five minutes to weekly, and writes the rows into your Airtable table. Match rows by a key so edits land in place and rows that left the query are removed, or simply append. No Zapier, no scripts, no exports. About fifteen minutes the first time, and the Free plan covers it.

About 15 minutes No coding needed File stays where it is Every 5 minutes to weekly

What you’ll end up with

An Airtable table that holds the rows your DuckDB query returns and is brought up to date on a schedule, every five minutes to weekly, with a receipt of what changed on every run. Like this:

Orders to Airtable Active

Daily sync · 09:00 · Airtable 1,643 rows · 8.8s

1,627rows in syncmatched by key · nothing to do

11newappended under your headers, same columns

5updatedprices changed in DuckDB, 5 rows updated in place

0removedrows leaving the query are removed from the table

DuckDB holds your data in a file. Your team works in Airtable. The usual bridge is a Zapier or Make scenario, or a CSV someone uploads that breaks quietly and copies one row at a time. Query Streams Automation does it differently: the Agent that already reads your DuckDB file runs your saved SQL on a schedule and writes the whole result into Airtable in one pass, matching rows by a key so the table stays correct rather than growing.

What you’ll need

A Query Streams account

The Free plan is enough for this guide. Sign up at my.querystreams.com.

DuckDB connected through the Agent

The Agent installed on the computer that holds your DuckDB file, or on one that reads it from a network share, with the connection added in Data Agents. Not done yet? Connect DuckDB first; it takes about ten minutes.

A saved query

The SELECT you want to keep in sync, saved in Query Builder. Step 1 below shows one. If you want changed rows updated in place, include a column that is unique per row, like an ID.

Your Airtable login

Airtable asks you to let Query Streams read and write records, and read and write table structure, in the bases you choose. A Personal Access Token works too if your workspace prefers it.

Set it up in six steps

1. Save the query you want to keep in sync

   Open Query Builder, pick your DuckDB connection (not connected yet? start here), and write the SELECT that returns exactly the rows and columns you want in Airtable. This one pulls a few columns from an orders table:

   Example · SQL

   SELECT order_id, customer_name, order_total
   FROM main.orders
   ORDER BY order_total DESC
   LIMIT 100;

   Run it, then Save it with a name you will recognise, like “Orders for Airtable”. Keep a column that is unique per row in the SELECT, here order_id. That is the key Full Sync uses to update rows in place. Edit and re-save the query later and the automation picks up the change on its next run.

2. Open Automation and choose Airtable

   In the portal, open Automation (top menu Streams → Scheduled → Automation) and click New automation. The wizard has five short steps. On Step 1 click the Airtable card.

   Open Automation

   Seeing “Before You Begin” instead? The wizard only opens once an Agent is online with a connection and at least one saved query exists. The page lists which of the three is missing and links to it.

3. Connect Airtable

   On Step 2 click Connect with Airtable. Airtable opens, asks which bases Query Streams may use, and sends you back. Airtable asks you to let Query Streams read and write records, and read and write table structure, in the bases you choose. A Personal Access Token works too if your workspace prefers it. Already connected once? The page offers Use This Connection instead.

4. Choose where the rows land

   On Step 3 pick the Base, then the Table the rows should land in. Starting fresh is easiest: choose Create New Table, type a name and click Create; Query Streams creates the fields from your query’s columns. If the table already has rows, a preview shows them and explains what each sync mode would do to them. One automation per table.

5. Pick the query and decide how rows land

   On Step 4 open the Saved Query list and choose the query from step 1. Query Streams runs it once to read the columns, tells you how many rows it returns, and shows two choices:

   Append · fast

   Append

   Every run adds the query’s rows to the bottom of the table. Nothing is updated and nothing is deleted, so it is the quick choice for logs, events and history you want to accumulate. No key column needed. A row that comes back unchanged is not added again; a row whose values changed is added below as a new line and the old one stays.

   Full Sync · mirrors

   Full Sync

   Rows are matched by a key column such as an order ID. New rows are added, changed rows are updated in place, and rows that no longer come back from the query are removed, including rows someone typed into the table by hand, so keep hand-made rows on another table. This is the choice for inventory, pipelines, master data and dashboards.

   Full Sync is set per column, which is what makes it useful in a shared table:

   Key

   The column that identifies a row, for example order_id. It has to be unique in the query result; Query Streams checks that for you and refuses duplicates.

   Sync

   Overwritten from DuckDB on every run. If someone edits a Sync column in Airtable, the next run puts the source value back.

   Preserve

   Written when the row first arrives and never compared again, so a change to it in DuckDB does not trigger an update. For notes, owners and follow-up dates your team types by hand, use a column that is not in the query at all: columns that exist only in Airtable are never touched by a sync.

   Which is faster? Append writes without reading the table first. Full Sync reads it, compares every row, and writes only the differences, so it takes longer but the table always matches the query. You can switch later from the automation’s Edit panel.

   For a table you want to trust, click Create all columns as Sync, then click the key icon on order_id to make it the Key. Query Streams scans the whole result to confirm the key is unique and warns you if it is not. Keep your team’s own notes in columns that are not part of the query; those are never touched. For a log that should only ever grow, click Create all columns as Preserve instead and skip the key. Click Continue.

6. Set the schedule, name it, create it

   On Step 5 choose Run on a schedule and pick a Frequency:

   Frequency	Runs
   Every 5 Minutes	every 5 minutes, the fastest allowed
   Every 15 Minutes	every 15 minutes
   Every 30 Minutes	every 30 minutes
   Every Hour	at the top of every hour
   Daily at 9 AM	every day at 9:00, the default
   Daily at 6 AM	every day at 6:00
   Weekdays at 9 AM	Monday to Friday at 9:00
   Weekly (Monday 9 AM)	every Monday at 9:00
   Custom Schedule	your own minutes, hours, days of the week or day of the month

   Times follow the Timezone you pick, which starts as your browser’s. You can also choose Manual only and press Run whenever you like; a schedule can be added later from the automation’s Edit panel. Scheduled runs happen on the Agent, so they keep going while your browser is closed. If the computer running the Agent is off when a run is due, the missed run happens once, shortly after the Agent starts again.

   Give it an Integration Name, such as “Orders to Airtable, daily”, check the Ready to Create summary and click Create Automation. Back on the Automation page, press Run once to do the first sync now rather than waiting for the schedule. Open Airtable: the rows are there, under your column headers.

What you see afterwards

The Automation page lists every automation with its status, source, target, key, schedule, last sync and row count. Open one for the Executions tab: every run with its trigger (manual or scheduled), duration, and rows created, updated, deleted and unchanged. A Performance tab shows data processed and success rate over 7 to 90 days.

What the plans allow

Automation is on every plan, including Free, with no limit on how many automations you create. What each plan limits is how many rows one query may return per run:

If a query returns more, the run is held rather than cut short, and the automation tells you which plan fits or suggests adding a filter such as a date range. Each successful run also counts against your Data Realm allowance, the same pool your Excel and Sheets queries use. See the pricing page for the allowances.

Good next moves

• Add a filter to the query, such as the last 90 days, so the table stays small and fast and fits your plan’s row limit.
• Keep hand-typed notes in their own columns, outside the query, so they survive every sync.
• Group your automations from the Edit panel once you have a few; the Automation page filters by group.

Reaching a file on a network share

The Agent opens files as a background service, not as you, so a file on another computer needs one extra step. Which step depends on the computer the Agent runs on.

If the test cannot get to the share, look for these words in its result: unreachable means the Agent’s computer got no answer from the server. On Linux or a Mac, not mounted means the share still has to be mounted, and that message includes the exact mount command for your server.

Test failed on a share?

Every run reaches the file the same way the connection test does. If runs start failing after something changed on the file server, a new password for example, open the connector in Data Agents and click Test connection. The test result, and the line under the file path, name the problem:

Questions people ask

Related guides

• MySQL to Airtable
• PostgreSQL to Airtable
• Microsoft SQL Server to Airtable
• DuckDB to Google Sheets
• DuckDB to Smartsheet
• How to connect DuckDB to Google Sheets
• All Airtable automation guides