Check whether table headers sit on the same line

Where the extra work starts

Test the heading and one complete row together so you can see whether the layout needs correction before processing a larger batch.

ParseMill processes native invoice PDFs and clear printed English PDF scans. Try a typical supplier invoice and a difficult one. Check the extracted fields, correct exceptions and open the workbook before committing to a monthly plan.

ParseMill's invoice workflow recognizes supported headings, including multiword labels such as Unit Price, within horizontal tables. Staggered and multi-level arrangements still need a specific fit test. Review the resulting field associations instead of assuming that a visually recognizable label guarantees correct mapping.

A practical way to get it right

Zoom into the heading region and compare the baselines of the required labels. Note whether a word is a group heading, a unit below a heading or a continuation of a multiword label. Copy the text as a character check, but retain the visual position because a flat paste may hide the vertical differences.

Record the heading sequence and position in a small fit worksheet. Distinguish aligned labels from staggered labels and multi-level groups. Then inspect the first data row to ensure it is not being mistaken for a missing heading.

If the arrangement falls outside the supported shape, stop the automatic attempt for that layout and choose a reviewed alternative. Do not add fake headings to the source PDF or rename data cells to make the test appear to pass. A manually prepared derivative needs its own provenance and verification.

Adobe: Copy content from PDFs describes text selection and copying and how to check document restrictions. Use copying to check characters, while separately inspecting the visual baselines that a plain paste does not preserve reliably.

What this looks like on a real task

Synthetic example. The people, references and figures below are invented to demonstrate the method.

The first row passes only this heading-position check. It still needs readable text, suitable rows and source review. Copy the arrangement and fit-decision columns and add a source page reference. Keep failed heading shapes in the sample set rather than replacing them with an ideal reconstruction.

Catch the cases that create rework

A title above the table may contain one of the expected words and create a misleading apparent match. Confirm that each label governs the intended data column. Likewise, units below headings may need preservation even if they are not configured as columns.

If another tool promotes the first row automatically, inspect whether it selected the true heading row or a data row. An output column named 12.50 is a clue to investigate, not a meaningful schema.

Microsoft: Promote or demote column headers explains that header promotion can misidentify the header row. Verify the chosen labels against the source rather than accepting automatic promotion as proof of fit.

Decide whether the change is worth keeping

The fit note should name the exact labels, their relative positions and the unsupported condition if present. Test only a suitable representative file and inspect the resulting source associations. The decision should be reproducible from the page rather than based on a general impression that the table looks similar to an earlier invoice.