If you want to analyse spending in a spreadsheet, choose CSV or Excel whenever your bank offers it: the rows are already in cells, so there is no conversion and far less to repair. Pick PDF when it is the only format on offer, or when you need an official copy to show someone.
How do the two formats compare?
| Point | CSV or Excel download | PDF download |
|---|---|---|
| Effort to get into a sheet | Open it, or import it | Convert first, then clean |
| Risk of lost or shifted rows | Low | Higher, depends on layout |
| Narrations in one cell | Usually yes | Often wrapped onto two lines |
| Looks official for visa or loan files | Not usually | Yes, this is its job |
| Print-ready with headings and summary | No | Yes |
| Dates and amounts as real values | Often text, needs a check | Text after conversion, needs a check |
Neither format fixes everything. A CSV can still hold dates as text, and a PDF can convert cleanly. But the typical sequence is: CSV needs a few minutes of tidying, PDF needs a conversion tool and then a longer cleanup.
Is a CSV always complete?
Not automatically. Some banks limit how many months or rows a download covers, or leave columns out. Whichever format you choose, open the file and compare three things against a short PDF or the app: the first date, the last date and the closing balance. If the CSV stops a week early, you have found out before building a year of analysis on it.
CSV traps to know about
- Double-clicking a CSV in Excel applies your computer's regional date order, so 05/01/2024 may become 1 May. Importing with Data, From Text/CSV lets you set the column type yourself.
- Long reference numbers can turn into scientific notation. Excel keeps only 15 significant digits, so a long number can lose its tail. Set that column to Text before loading.
- Leading zeros in cheque or reference numbers vanish if the column is read as numbers.
- Amounts with commas inside may arrive as text. Test with ISNUMBER before summing.
When only a PDF is available
That is common for older periods and for some card accounts. A workable rule of thumb:
- Try Excel's own PDF import first (Data, Get Data, From File, From PDF), where your version has it, and see how the table looks in the preview.
- If the preview shows one clean table per page, load it and clean the repeated headings.
- If narrations wrap or columns drift, expect more repair, and plan the checks: row count, totals and closing balance.
- If the statements are scanned images rather than text, the result will need typing or a character-recognition step, and the mistakes will be yours to catch.
Which should I choose for monthly tracking?
Say you want to track one account for twelve months. With CSV you might spend ten minutes per month tidying dates and a few columns. With PDF the same month might need a conversion, a heading cleanup, a narration merge and a balance check, easily four or five times as long. Over a year, that is the difference between a habit that survives and one that you drop in March.
Hypothetical times, of course. Time one month of each and decide from your own numbers.
A last practical point: if you track several accounts, keep the formats consistent where you can. Mixing a CSV from one bank and a converted PDF from another means two sets of date rules and two sets of checks, so write down, per account, which format you use and what to verify each month.
A short checklist
- Look at your bank's statement download page and see which formats exist for the period you need.
- Prefer CSV or Excel for analysis, PDF for proof.
- Verify first date, last date and closing balance after loading.
- If you would rather not clean any of it, SpendixAI accepts PDF, CSV and Excel statements and shows the rows for you to review before saving.
Useful ideas start with a clearer picture of your own spending.
Explore your reports

