Check a rotated PDF table before choosing a workflow

Where the extra work starts

Try the page in its actual orientation and check the result so a readable PDF is not mistaken for a supported extraction format.

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.

Do not infer ParseMill support from a viewer's ability to turn the page. Rotated-table behavior has to be established on the specific input and current workflow; an untested orientation should remain an explicit fit uncertainty.

A practical way to get it right

Record the source page orientation and whether the text itself is rotated within the page. Use the viewer's rotation only to inspect the content initially. Select and copy a heading and row to assess characters and sequence, then note the table's column relationships.

If an authorized derivative is prepared with persistent page rotation, save it separately and retain the original. Reopen the saved derivative to confirm the intended orientation persisted. Record the preparation step and compare headings, rows and values so the transformation does not conceal a missing region.

Choose a small bounded test only if the resulting file otherwise fits the supported structure. Compare all candidate fields and source associations. If behavior is untested or output is unreliable, use reviewed manual transcription or another suitable route and record the limitation.

Adobe: Adjusting PDF views states that Rotate View changes only the display and cannot be saved; persistent rotation uses Organize pages. Record which operation actually affected the tested file.

What this looks like on a real task

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

The derivative is not declared supported merely because it opens upright. Copy the observation, result and decision columns and add the actual trial evidence if a test is performed. This example deliberately ends at unconfirmed fit rather than inventing a result for a rotated document.

Catch the cases that create rework

Temporary rotation can disappear when the document is reopened or uploaded elsewhere. Do not assume the parser sees the same view you selected. Text may also be rotated inside an otherwise upright page, so page orientation alone may not resolve the geometry.

A copied row may look correct while its source highlights point to the wrong positions. Verify the evidence relationship as well as the exported strings during any trial.

Adobe: Copy content from PDFs describes text selection and copying and how to check document restrictions. Use it as a character check while separately verifying orientation-dependent source associations in the extraction result.

Decide whether the change is worth keeping

The final record should identify the original and derivative files, preparation performed, tested pages and observed outcome. Accept the method only when the current result is verified against the source. Keep the rotated fixture in the acceptance set if reuse is planned, and repeat the relevant check when the parser or preparation route changes.