How to train Azure Form Recognizer to recognize a optional field?

Question

I'm trying to train a model, but, one field of my dataset is optional. When analyzing the document, the model doesnt return the expected result.

the blue field represents the field that I'm trying to map. In document 1, it is not present, so, the return should be empty.

in document 2, the field is present and the model is able to retrieve the data.

when the value is not present, the model maps to another random field, for example "Nome Fantasia", instead of return empty.

how can I tell Azure Form Recognizer that the field may or may not be present?

or, second question, how can I inform model that the value is aways below "Data de fim da sanção" label?

Answer 1

Yes, you can use OCR to extract and label data it works with pre-trained models that can recognize the key-value pairs, text, and tables from your documents and the table contents in the file uploaded as the input. Even though paragraphs and table content in the middle ,file contains a large amount of the text or at any place, it will be recognized.

labelling OCR files refer to this Link.

I reproduce a similar issue with Form recognizer refer to this SO thread and github link.