BankXL / Practical guide

Bank statement PDF to Excel: build a clean worksheet from text pages

Use Excel or a local PDF extractor, then repair page boundaries and data types before calculating statement activity.

Try the spreadsheet route first

To move a bank statement PDF to Excel, first ask whether a transaction CSV is available from the bank. If it is, import that instead and use the PDF for checking. If you only have a text PDF and your Excel version offers the PDF connector, use Data, Get Data, From File, From PDF. Select a detected table in Navigator and choose Transform Data before loading. Microsoft lists the PDF connector for Microsoft 365 subscriptions of Excel for Windows, not for Office 2016 or 2019 or Excel for Mac, so a missing menu item does not mean your PDF is invalid.

Select transactions rather than the whole document

In Power Query, inspect each candidate table against its source page. Statement summaries, contact details and transaction activity can be detected as separate tables. Retain date, description, amount and running balance where present. Remove repeated column headings only after confirming that they are headings. A blank date may belong to a wrapped description rather than a new transaction. Do not apply Fill Down to an entire date column without checking how continuation lines are represented, because that can manufacture extra transactions from explanatory text.

Repair page boundaries explicitly

Check the final transaction on one page and the first transaction on the next. A brought-forward balance is context, not new activity. A description can continue onto another page, and a new page may switch spacing while retaining the same logical columns. Add a source-page column in your working sheet if you need to track repairs. When combining extracted tables, align columns by their meaning rather than their position. Power Query's PDF connector combines similar tables on consecutive pages by default, so check a combined table at every page boundary for repeated headers or shifted columns. A blank debit column on one page must not shift credits into the wrong field.

Protect dates and identifiers before loading

Remove or adjust automatic type-change steps when they interpret printed text incorrectly. A reference such as 00123 should remain text; an ambiguous date such as 04/05 needs the statement's locale and period. Keep a copy of the original printed date before making a normalized date column. Amounts need the decimal convention confirmed from the document, not guessed from one value. Check parentheses, trailing minus signs and separate debit/credit headings. A worksheet displaying a currency symbol has not necessarily stored the amount as a usable number.

Validate the finished worksheet

Count transactions for every source page, compare dates at both ends of the period, and total deposits and withdrawals separately. Then compare opening balance plus activity with closing balance. Investigate any difference from the original PDF; do not adjust an unexplained amount merely to force a match. If Excel does not produce a usable table, Tabula can export a selected text-PDF region to CSV for free. Import that CSV with explicit text and number types. Scanned pages require recognition or manual entry before these table-cleaning steps can work.

Make a repair table for page joins

For a statement spanning three pages, write down the last dated transaction on page one and the first dated transaction on page two before appending tables. Repeat at the next boundary. Keep an extra column describing whether a row is activity, a repeated heading, a carry-forward or a description continuation. Filter by that reviewed classification rather than deleting every row with an empty date. If a transaction amount appears on one line and its description on another, reconstruct one record from both source lines and retain the source page. This gives you a visible explanation for a row-count reduction during cleanup.

Build calculations beside preserved source text

Suppose the PDF prints separate money-out and money-in fields. Retain both original columns and add a numeric Signed amount column after verifying which side increases the account. A withdrawal of 35.20 with no deposit becomes -35.20; a deposit of 200.00 with no withdrawal becomes 200.00. Keep the running balance outside that calculation. Flag any row with amounts on both sides for source review rather than choosing one silently. Finally, filter the numeric column for errors and blanks before summing: a sum that ignores text cells can look convincing while omitting an unconverted amount.

Where BankXL fits

BankXL can supply a reviewable transaction table when local PDF import is unsuitable. It preserves printed dates for you to interpret, supports corrections, and exports XLSX with a Balance check sheet or CSV. It does not infer accounting categories or import into a ledger. Limits per conversion are one PDF, 15 MiB, 100 pages and 3,000 transactions; decimal-comma amounts and encrypted PDFs are unsupported. Trial output covers the first two data pages found within the first six pages. Continuing the same trial file keeps your saved results and corrections and charges only additional successful pages; its original expiry remains unchanged.

Check two pages with data free See page packs

Written by Zijian Xiao. Examples are synthetic. Review all extracted results against your original PDF.