Release Notes

June 2026

Our Files connectors now support creating empty columns and columns with null values in your destination. During the initial sync, we create these empty columns with the STRING data type. However, if the column later contains non-null values in subsequent syncs, we update the column data type based on the data type we infer from the schema. Previously, we did not support creating empty columns or columns with null values in your destination.

We support this enhancement only for the following File connectors set up using BigQuery, Databricks, and Snowflake as destinations:

We are gradually rolling out this enhancement to all existing connections.

May 2026

We have discontinued the archive_pattern and pattern parameters in the Fivetran REST API for the following connectors:

Amazon S3
Azure Blob Storage
Box
Dropbox
Email
FTP
Google Cloud Storage
Google Drive
SharePoint
SFTP
S3-Compatible Storage

We will continue supporting these parameters for both new and existing connections until August 31, 2026. After August 31, 2026, we will stop supporting these parameters for all connections and replace them with the files.archive_pattern and files.pattern parameters respectively.

April 2026

We now allow you to update the primary key column after the initial sync when using the Upsert file using custom primary key option.

We have added a new toggle, Allow inconsistent row lengths, to the connection setup forms to support syncing rows in CSV files with inconsistent lengths for the following connectors:

Amazon S3
Azure Blob Storage
Box
Dropbox
Email
FTP
Google Cloud Storage
Google Drive
SharePoint
SFTP
S3-Compatible Storage

This feature is available only to all new connections created after 13 April, 2026. We are gradually rolling out this feature to all existing connections. For more information, see our documentation.

We have modified how the _line system column reflects the row position in your destination. The _line system column now displays its physical row position in the source file, even if we skip rows from the sync due to the on_error=skip setting in your destination. Therefore, if we now skip syncing a physical row, the position of the _line column in your destination will remain unchanged. It will always match the corresponding row positions in the source file to ensure a consistent mapping between the source file and your destination. Previously, we updated the _line column sequentially, which was calculated based on the rows that were successfully synced. This change is available only for new connections created after April 22, 2026.

February 2026

S3-compatible Storage is a storage system that supports the standard Amazon S3 REST API. Fivetran supports syncing files from your S3-Compatible Storage bucket to your destination. Read our S3-Compatible Storage connector documentation.