---
title: "JSON data type is now available"
excerpt: "JSON data type, improved file matching for S3 connector, and much more!"
date: "2024-11-22"
---

# JSON is now available as a data type

You can now use the `JSON` data type in your Data Sources. See [Supported data types](/sql-reference/data-types).

For example, you can create an NDJSON Data Source using the JSON column type through the schema parameter, like this:

```shell
curl \\
            -H "Authorization: Bearer <DATASOURCES:CREATE token>" \\
            -X POST "https://api.tinybird.co/v0/datasources" \\
            -d "name=example" \\
            -d "format=ndjson" \\
            -d "schema=data JSON `json:$`"
```

{% callout type="caution" %}
As the JSON data type is being rolled out to all regions and clusters, it might not be immediately available in your Workspaces.
{% /callout %}

## Improved file matching for S3 Connector

From now on, new Data Sources that rely on the S3 Connector can benefit from improved pattern matching for files, including support for glob-like patterns and proper handling of `**` and `*` wildcards.

Data Sources using the previous patterns will continue working as usual. For more information, see [S3 file URI](/forward/ingest-data/connectors/s3#s3-file-uri).

## Engine settings for S3 Data Sources

You can now set the engine settings of your S3 Data Sources when creating them from the UI.

{% image src="/assets/changelog/s3-engine-settings.png" alt="Engine settings configuration for S3 Data Sources" /%}

## Other improvements

- Added the 12 hours quick option in Time Series.
- Implemented retries in the DynamoDB connector. Avoids data loss in case of downtime.
- Added `CPUUsage` metric to help customers to monitor their cluster CPU usage.
- `tb branch ls` returns two new fields, `created_at` and `owner`. Use them to checke whether a branch is active.
- Downstream materializations caused by failed ingestions now generate `datasources_ops_log` entries.

## Bugfixes

- Chart endpoints can now handle request URLs with a large number of parameters.
- Fixed an issue where x-axis values in stacked charts were being displayed multiple times.
- Fixed a 404 error when trying to redirect users after reaching a workspace without access.
- Fixed corner case in the retry strategy of the Kafka connector.