Can someone explain what “Edit on screen” does and where documentation exists describing it and/or something that shows how to use it in a Youtube (or otherwise) video ?
Also curious if “Edit on screen” has any special functionality for CV activities… like indicating columns that were misinterpreted ?
This field actually takes you back to the screen you used to introduce the table.
1-You can add a new column.
2-If your table contains more than one page, you can define a next button.
3-You can edit for each column. (name, type, etc.)
4-You can delete an existing column.
5-You can control the limit of the data you will extract.
6-You can check the extracted data.
In this area you can check, remove, and re-create incorrectly interpreted columns.
The Modern Experience Extract Table feature in Studio does not appear to work in PDF documents. When I attempt to identity “the table” within the PDF I get an error indicating this:
This element is not eligible for table extraction. Please select elements which are part of a table.
I’m looking to see if there is another way to pull table-like data from a PDF where Document Understanding has trouble with it.
Form Extractor works but the table “size” across tables varies in rows. DU ML Extractor also works but has its own challenges with the following:
The issue is the spaces which OCR at times assumes are variable font vs a fixed font. Once the data is read in, the columnar structure of the data is lost typically as “spaces” are lost essentially.
I was hoping to see if CV Extract Data works and it sorta works but again it has issues with the columns. It can’t find the columns and assumes correctly where they are (for many yes but others no it get it wrong) and you apparently can not “adjust” where the columns start/end or add columns that I can find.