How should we label the table that spans multiple pages?

Hi @Hrushikesh Shinde

Thank you for using the Microsoft Q&A forum.

Tabular fields in Document intelligence support cross page tables by default. To label a table that spans multiple pages, you need to label each row of the table across the different pages in a single table. As a best practice, it's recommended to ensure that your dataset contains a few samples of the expected variations. Please follow this documentation: https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/concept-custom-label?view=doc-intel-4.0.0#tabular-fields

Multi page tables: When tables span multiple pages, label a single table. Add documents to the training dataset with the expected variations represented—documents with the table on a single page only and documents with the table spanning two or more pages with all the rows labeled. https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/concept-custom-label?view=doc-intel-4.0.0#create-a-balanced-dataset

you can create table tags by selecting Add a new table tag. Specify whether the table has a fixed number of rows or variable number of rows depending on the document and define the schema.

In this kind of cross-page table scenarios try with Custom Neural model for best results.

I hope this helps. Thank you.

Hi @Hrushikesh Shinde

Following up to see if the below answer was helpful.

hey thanks for the response ..but what i mean when i say table that spans multiple pages is the single cell that got splitted across the multiple pages below is the reference image in this image you can see that content of the single cell is got splitted acorss two pages so how can we label the data in such a cases. thanks again

Hi @Hrushikesh Shinde

Thank you for giving more clarity, that the text spans across the cell within the same row of the table.

Yes, this is a known behavior. If you are trying to label the multi-page document data under the same field this cross-page label error is thrown.

Currently, this is not supported. Fields are considered global to the whole document and must be named uniquely. We don't have any ETA regarding this to share with you at this moment.

However, you can create separate similar labels (example: Field1, Field2) from the multi-page document. With this label practice you can extract the data even from the multi-page document.

You can also split the document and upload as a separate document and label. Please note that based on your use case and convenience you can opt.

However, we appreciate your feedback and encourage you to submit a feature request for "Cross-Page Labeling in Document Intelligence" through the "Share your idea" Azure Forum. Here's the link: Post idea · Community (azure.com).

I hope you understand.