Enable Automatic Handling or Configurable Conversion of PostgreSQL NUMERIC Columns in Copy Activity
When ingesting data from a PostgreSQL database into Microsoft Fabric using the Copy Activity, the process currently fails if a source column contains NUMERIC values that exceed the range supported by the .NET System.Decimal type.
During ingestion into OneLake (Parquet), the pipeline returns the following error:
Failure happened on 'Source' side. ErrorCode=UserErrorInvalidDataValue… Type=System.OverflowException, Message=Numeric value does not fit in a System.Decimal.
This behavior requires users to manually convert each PostgreSQL NUMERIC column to a double type before ingestion. While this workaround succeeds technically, it is not scalable or maintainable—especially for large tables, dynamic schemas, or environments with frequent schema changes.
Request
I kindly request an enhancement to Microsoft Fabric that enables seamless ingestion of PostgreSQL NUMERIC columns without requiring manual pre-processing. Specifically, the solution could take one of the following forms:
Automatic handling of wide-range NUMERIC values
Automatically map PostgreSQL NUMERIC values to a suitable type (such as double, decimal with expanded precision, or string) when they exceed the .NET Decimal limits.
Configurable type conversion in Copy Activity
Provide a setting that allows users to specify how NUMERIC columns should be handled during ingestion (e.g., convert to double, string, or preserve as-is with higher precision).
Benefits
Eliminates manual column-by-column conversion
Reduces ingestion failures caused by numeric overflow
Improves scalability for large or frequently changing PostgreSQL schemas
Enhances user experience and reliability of Fabric ingestion pipelines
This feature would significantly streamline PostgreSQL ingestion into OneLake and help users avoid unnecessary operational
overhead.
Thank you for considering this improvement.
3 Comments
- AnonymousNot applicableHello Team, This is one of Microsoft's customer requests as the numeric column for the Postgres sql is having limited exception to store the data. Customer is expecting to increase the size to store the data because it is blocking them. please consider this feature request and please try to add this feature soon. Thank you!
- AnonymousNot applicableDear Team, One of Microsoft’s customers has reported a limitation with the numeric column in PostgreSQL, which is restricting their ability to store data. They are requesting an enhancement to increase the column size to accommodate larger values, as this limitation is currently blocking their operations. Please consider prioritizing this feature request and adding the capability at the earliest possible opportunity. Thank you for your support!
- andyclapAdvocate IV+1 - it's annoying especially as the exception only triggers on numerics when it finds a value that is out of range. Mostly it's just over-precise rounding for us, would be nice to be able to map types client side.
Recent ideas
Data Pipelines - Run only selected activities
For debugging and testing pipeline activities during development, allow us to select one or multiple activities and run only the selected pipeline activities. For example, I'm working on editing ...frithjof_v8 hours agoCommunity ChampionNew612Views11likes2CommentsSemantic model connection bindings should be in source control (Git)
Semantic model data source connection bindings should be source controlled. A semantic model can contain multiple data source references, each of which can be mapped to a separate Fabric data connec...frithjof_v15 hours agoCommunity ChampionNew17Views1like0CommentsBulk changing column names in Visualizations Pane
We often use raw/api column names or measures with a set nomenclature to be consistent and to keep track of them but we do not want to display these names in the visuals. Currently we have to change ...vishal14019719 hours agoFrequent VisitorNew6Views0likes0Comments