I have a dynamic workflow that takes files and loads them into postGIS using FeatureWriter. I am having issues with datasets that contain date fields with some null values resulting in features getting skipped.

The writer seems to be trying to write '00000000' for these features rather than leaving it as null

Is there a way to stop this behaviour? In a standard workbench I could set a default value but it feels like I shouldn't need to do that as postgres can handle null values in date fields and I don't know how I could change the values as it's a dynamic schema.

Errors in log:

Value of attribute 'createdate' could not be converted to type 'date'. Feature will be logged and skipped

Attribute(string: UTF-8)      : `CREATEDATE' has value `00000000'

---

Schema definition for field:

attribute{8}.fme_data_type fme_date

attribute{8}.name CREATEDATE

attribute{8}.native_data_type date

The next line is missing for features that do not have a value:

CREATEDATE 20031118