CSV Comma-Separated Decimal Values Sync as STRING or VARCHAR

Issue

When a CSV file contains a column with comma-separated decimal values, the column is created in the destination with the STRING or VARCHAR data type.

Environment

File connectors

Resolution

To resolve this issue, you must modify your source CSV files and how they're generated:

Find and replace all commas (,) used as decimal separators with periods (.).
Configure your source system to generate CSV files using periods as decimal separators.
After correcting all CSV files, drop the affected destination table and manually trigger a historical re-sync for your connection. For more information, see How to Trigger Historical Re-Syncs for Fivetran Connections and Tables.

Once a column has been created in the destination with the STRING data type, we don't update the data type. You must drop and re-create the table for the updated data type to take effect.

Cause

The columns are being synced as strings because the decimal separator needs to be a decimal point, or period, instead of a comma.