Field note_nlp."offset" should be integer, not varchar

[ekorchmar]

From the documentation

Character offset of the extracted term in the input note

I am assuming that in this context this field is expected to contain a numeric value: index position of start character of the extracted token (note_nlp.lexical_variant) in the string of the entire text (note.note_text).

For instance, if the note text is "Diagnosed with eczema", value of note_nlp.lexical_variant would be eczema and value of note_nlp."offset" would be 15.

My assumption is that the purpose of this field is to specify which occurence of the same literal token is referred to. So it should just be an integer, with no extra information.