How to create a new table in an existing base through the API?

[Dev_Local]

I have searched the API docs, and I see no way to do this.

With the AirTable API, how do I create a new tab (new table) in an existing base? and then populate it?

I do not see this feature of inserting or creating an new tab into an existing base, only the ability to insert rows into an existing table that was created by a human that opens a web browser and clicks.

We have needs to automate the creation of new tables in an existing base (and down the line, creating new bases programatically).

[EvanHahn]

:wave: Developer on Airtable’s API team here.

The Airtable API currently has no way to do this—you can only create new tables from the UI.

[Dev_Local]

Great, thank you. Is there any way with the AirTable API, or any 3rd party product to insert a CSV direct into an existing base?

I have a CSV on a file system, and I want to programatically send a CSV to a base and have it inserted into a new tab.

I can’t be the first to request this, it would be helpful, if at all possible. It does not look possible with Zapier. Thank You.

[EvanHahn]

There’s no way to import a CSV directly with the API, but you can convert CSV to JSON, which our API accepts.

So rather than CSV → Airtable, you’d do CSV → JSON → Airtable.

Does that make sense?

[Dev_Local]

Yes, so where can I find documentation on that? When I goto my Base’s API page, I don’t see anything there, will look again.

And so sending a JSON direct, would this create a new tab in the base? and also do I have control over the naming of that tab? (or positioning?)

thank you,

[EvanHahn]

Dev_Local:

Yes, so where can I find documentation on that?

We don’t document that because it’s a bit out of scope, and depends on what exactly you’re trying to do. A quick search shows a few online tools that do this conversion, and you can also look to see how to do it in your favorite programming language (like Python).

Dev_Local:

And so sending a JSON direct, would this create a new tab in the base? and also do I have control over the naming of that tab? (or positioning?)

There’s a new tab for each table in the base. Again, you can’t create a new table from the API.

[Dev_Local]

So the solution is that I have to create a new table manually, always, with the headers I want, and once the table exists, with a header for each field I want to insert, I can then I can make API calls to insert data.

I can send up to ten records at a time, per API call, and I can make up to five of these calls per second, meaning a max of fifty records inserted per second into an existing table.

If I got any of that wrong, please let me know, this is what I am reading in the documentation.

---

I understand how to take my data and convert to JSON.

I was just hoping that I could create a JSON file of 1K records (rows), and make one call and insert all 1K into a base, creating a new table at the same time and setting the tab name.

Thank you,

[EvanHahn]

All of that understanding sounds correct.

We may add support for these features in the API in the future, but they aren’t currently available.

[Bill_French]

Dev_Local:

… was just hoping that I could create a JSON file of 1K records (rows), and make one call and insert all 1K into a base, creating a new table at the same time and setting the tab name.

Um, yeah - me too. This is a common process for anything automated. It tends to rule out a number of use cases.

[JonathanBowen]

Here’s a Python script to import a CSV (turning into a JSON object before posting to AT):

How to programatically import CSV to AirTable, creating a new table in an existing base (CSV is a fi... API

For those that are interested, here’s an example Python script that imports records one by one (takes only a few seconds for 100 records): import requests import csv post_url = 'https://api.airtable.com/v0/YOUR_APP_ID/CSVImport' post_headers = { 'Authorization' : 'Bearer YOUR API KEY', 'Content-Type': 'application/json' } f = open('MOCK_DATA.csv') csv_f = csv.reader(f) for row in csv_f: name = row[0] number = int(row[1]) role = row[2] data = { "fields": { …

JB