Bank statement columns shift at a page break: check the boundary

When the second page suddenly has strange signs or balances, inspect its column mapping before changing values. A blank column or repeated heading can shift extracted cells. Review each page independently, then combine only matching fields.

Compare the rows on both sides of the page break

1. Compare the last two transactions of one page with the first two of the next. Read the printed column headings again.

2. Give each extracted page its own source-page field. Map date, description, debit, credit and balance by the source layout, not spreadsheet position alone.

3. Test the corrected boundary before appending the remaining rows. Keep the original extraction so you can reverse the mapping.

Fictional example

A fictional page uses Date | Text | Debit | Credit | Balance. The next page contains an extra blank column before Debit. A positional append puts debit 14.00 into Credit, reversing the movement. Mapping by confirmed headings restores a debit of 14.00 without editing the amount.

Prove the mapping with two opposite movements

Extend the fictional example with a last balance of 200.00 on page 1. Page 2 prints a debit of 14.00 and then a credit of 9.00, with balances 186.00 and 195.00. Check the complete pairs: 200.00 − 14.00 = 186.00, then 186.00 + 9.00 = 195.00. If both columns are reversed, the table instead gives 214.00 then 205.00. The final difference of 10.00 is not a missing fee to add: it is a reason to inspect the two assignments. Confirm the headings and printed rows before accepting this diagnosis.

Write a small mapping record for each layout: page, extracted column, printed heading, destination field, and one checked transaction. For page 2, an extra blank field must not silently become Debit. Apply the confirmed mapping to that layout only; the next page may change again. If the PDF has no running balances, use the printed debit/credit positions as evidence and keep the arithmetic check unavailable rather than inventing a balance.

What an append operation can and cannot repair

Match reviewed field meanings and column names before appending page tables. Appending tables cannot correct a wrongly named Debit column. Rename verified fields before appending. Inspect a separate working copy of the problem page before combining its reviewed rows with the rest. On a manual copy, keep separate page tables until this check passes.

Isolate page 2 and write its column map

In a copy of a Power Query query, the following M expression requests only page 2 of a fictional local PDF. Replace the example path with the absolute path of your own authorized local file. It is a documented option example, not an extraction tested by BankStatementScan.

Pdf.Tables(
    File.Contents("C:\Example\statement.pdf"),
    [StartPage = 2, EndPage = 2, MultiPageTables = false]
)

Record the tested page range and whether tables were combined before inspecting the returned rows. Selecting a page range does not guarantee one detected table per page. Inspect the returned tables and select the one containing the printed transactions. For the fictional extra-column layout, record this mapping before renaming fields:

Extracted field on page 2	Verified content in the PDF	Destination
Column1	Booking date	Date
Column2	Description	Description
Column3	Added empty layout column	Exclude after checking this layout
Column4	Debit 14.00	Debit
Column5	Credit position, empty for this debit	Credit; preserve the empty cell
Column6	Running balance 186.00	Balance

Do not shift every nonempty cell left: that would erase the distinction between an empty Credit cell and an extra layout column. Check the credit of 9.00 as well, since its Debit cell should remain empty. A field missing entirely from a page is different from an empty cell in an identified field. If the expected column cannot be located, hold that page for review rather than fill it with zeros.

Return to the step before any value replacement or type conversion that obscured the columns. Apply the confirmed names, then append a copy and compare row counts, debit totals and credit totals with the separate page tables. If the layout changes within a page, split it into verified sections instead of applying one shift to the entire page.

Checks and limits

Check incoming and outgoing rows on both sides of every layout change. If using Power Query’s PDF connector, its page-range options can help isolate the affected page; they do not validate the mapping. Do not use a balance discrepancy alone to decide which column to shift.

Related checks

Read column order and surrounding lines together at each page break. Bank PDF copy-paste in the wrong order: diagnose before sorting · Bank statement descriptions split across rows: join them safely · Diagnose a bank PDF before choosing a table extraction tool.

BankStatementScan is in preparation. The site offers an access request and a fictional example; it does not currently convert your bank files. See the planned service.

Sources consulted on 16 October 2026

• W3C — Model for Tabular Data and Metadata on the Web.
• W3C — PDF reading order.