When you configure an update, upsert, or delete operation in a Snowflake dynamic mapping, the Snowflake target table must have the primary key defined for the column. If you configure a mapping flow between the Snowflake source and target in the dynamic mapping, the keys selected in the Snowflake data object are not picked.

To rectify this, perform the following steps:

Add a transformation in the mapping.
To dynamically get the data object columns from the data source at runtime, select the
At runtime, get data object columns from the data source
check box.
Configure the mapping flow option between the source and the transformation, and then enable runtime linking between the transformation and the target.