Complex Tables Extraction

Note: this feature is currently available to Early Access applicants. If you wish to trial it, please email [email protected].

Extracting nested values in line items

When you are extracting "SKU" or "PO Number" for each line item on document you can experience the following situations:

  • The values are part of the "Description" column of the line item
  • "SKU" or "PO Number" is right above or below the line item

An example of such cases is visible on the table below.

614

One of the more common cases of semi-structured column

Extracting such values and linking them to the correct row automatically out of the box is challenging and even trying to capture the "Item Code" value manually with Magic grid is almost impossible. One would have to capture them one by one.

How Rossum helps with capturing nested values

To address this issue, we came up with an extension that focuses on making the manual part of the extraction as "automatic" as possible. We call the extension Magic items and once we would enable it on your account you can capture the complex tables in the following manner:

  1. Use Magic Grid for all the gridable columns
  2. Point and click on all the values in the first row that are non-gridable. In this case it is "PO Number" and "Item Code"
  3. Click on the "Extract complex line items" button
  4. All other values for the non-gridable columns should be captured
1842

Getting the best result out of this

In order to have the best extraction results, you have to capture all the lines of the table with a grid because Rossum is trying to find the nested values in the lines you created with the Magic grid. If you would disable some rows, Rossum would not look for the nested values there.

See the bright future

In the next couple of weeks and months we will focus on:

  • letting you accept the values suggested by the AI and thus giving you more control over the captured data
  • solving cases when one "PO Number" has to be distributed to multiple line items