How can I preserve or identify page breaks when a table spans multiple pages so that I can accurately associate the extracted content with its original page number?

Prafful Agarwal 20 Reputation points
2026-07-08T07:58:49.7833333+00:00

I am using Content Understanding to extract data from a document. There is a table that is split across two pages in the original document. However, in the generated Markdown output, Content Understanding merges the table into a single table and does not insert any page-break indicator between the pages.

Since I need to display the correct page number for each row/section, this causes issues because the page boundary information is lost during extraction. As a result, the page numbers shown are incorrect.

Azure Content Understanding in Foundry Tools
0 comments

1 answer

Sort by: Most helpful
Most helpful Newest Oldest
  1. Alex Burlachenko 25,120 Reputation points MVP Volunteer Moderator
    2026-07-08T09:58:25.1966667+00:00

    hi Prafful Agarwal & thx for sharing urs issue here at Q&A portal,

    If you're relying on the generated Markdown alone, I don't think there's a way to recover the original page break once the table has been merged.

    The Markdown output is meant to preserve the document structure, not necessarily the physical page layout. When a table spans pages, it's expected that it may be emitted as a single table without page-break markers.

    If you need row-level page attribution, I'd look at the structured output (JSON) instead of the Markdown. Check whether the extracted table cells, spans, or bounding regions include page references that you can map back to the source document.

    If the structured output doesn't expose page information for the table rows either, then this is currently a product limitation rather than something that can be reconstructed from the Markdown after the fact. I'd be interested to hear from the product team whether preserving page boundaries within multi-page tables is supported or on the roadmap.

    rgds,

    Alex

    &

    If my answer was helpful pls mark it and additional thx if u follow me at Q&A portal

    and at my blog https://ctrlaltdel.blog/

    Was this answer helpful?

    0 comments

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.