Date Format in Document Information Extraction

[carlonnheim]

Hi,

We are using Document Information Extraction and have a mix of date formats, such as "YYYY-MM-DD", "DD.MM.YYYY" and "MM.DD.YYYY" in the inputs.

The service guesses the right format if it is unambiguous, but misses on dates where exchanged order of day and month are still valid dates.

What is the approach to tell the Template and/or on uploaded document which date format to expect? The UI is not presenting any options to select this when preparing the template, and we are not finding any options to select locale on uploaded documents either?

Same question goes for documents uploaded on the REST API.

Thanks in advance!

//Carl

[tobias_weller]

Hi Carl,
Currently the service uses different aspects from the document to determine the correct data format. This also addresses cases in which multiple variations are valid. However, as you also described, there are cases left in which the detection yields the wrong result. We are planning on providing the option to define your date format, but we do not have a release scheduled for this. In the meantime, what you can do is create a ticket on our component (CA-ML-BDP) so that we can analyze why the detection in wrong in your documents.

Best regards,

Tobias

[Umesh_Pawar]

How can we handle date format discrepancies—specifically the US format (MM-DD-YYYY)—in SAP Document AI service when extracting invoice dates?