The cleanest way to get an ANZ statement as a CSV is to look for a spreadsheet or CSV download where you get the statement in the first place, and ask the bank if it is not offered. If all you have is a PDF, you can convert it in a spreadsheet program, but expect to tidy dates, amounts and columns by hand before the file is usable. Whichever route you take, check the opening and closing balance against the rows before relying on it.
Route one: ask for a CSV or Excel version first
A file produced directly from the bank's records beats any conversion, because nothing has to be guessed from a page layout. Look at the download options next to the statement, and check your own statement or the bank's help page for the formats on offer. If the choice is missing, ask the bank through the contact route printed on your card. Say which account, which months and which format you want.
Route two: convert a PDF in Excel or Google Sheets
- Open the PDF first to confirm it is not locked. If it asks for a password, sort that out before anything else.
- In Excel 365, look for the option to bring data in from a PDF under the Data tab. Pick the table the program finds, which is often named Table001 or similar. Older Excel versions do not have this feature.
- In Google Sheets, the PDF is not read directly. A common workaround is to copy the text from a PDF viewer and paste it in, then split it into columns with the Split text to columns option.
- Look at the first and last rows. Page headers and footers often repeat inside the data and need deleting.
- Save the finished sheet as CSV. In Excel, choose CSV UTF-8 so that rupee signs and accented names are kept.
Do not paste a statement into a random online converter. The file carries your account number and every merchant you paid.
What usually needs cleaning after conversion?
| Problem | What you see | Fix |
|---|---|---|
| Dates as text | Left-aligned, sort in the wrong order | Convert with the date-format options or retype one and fill down |
| Amounts as text | Sum comes out as 0 | Remove commas and currency signs, then change the column to number |
| Debit and credit in one column | Credits shown with a minus or a CR tag | Split into two columns or give credits a plus sign |
| Wrapped descriptions | One transaction over two rows | Join the rows and delete the spare |
| Repeated headers | Header row every page | Filter and delete them |
Indian number grouping such as 1,25,000.00 is read as text by many spreadsheets. After removing commas the value becomes 125000.
How do I know the CSV is right?
Use the statement's own totals as a check. Say the opening balance is ₹42,000, credits add up to ₹85,000 and debits to ₹61,350. The closing balance should be 42,000 + 85,000 − 61,350 = ₹65,650. If your converted rows give a different closing figure, a row is missing, doubled or has the wrong sign.
In a sheet, with credits in column D and debits in column E from row 2 to row 200, these formulas give the totals to compare:
- =SUM($D$2:$D$200) for credits
- =SUM($E$2:$E$200) for debits
- =42000+SUM($D$2:$D$200)-SUM($E$2:$E$200) for the closing balance, replacing 42000 with your own opening balance
When a CSV is more work than the job needs
If the real goal is to see where the money went, you may not need your own CSV at all. SpendixAI takes statements as PDF, CSV or Excel, so you can upload what the bank gave you and review the rows before they are saved. It does not promise that every layout parses perfectly, which is why the review step matters, and it never asks for net-banking logins.
Pick the account before importing, so each transaction keeps its bank or card. Upload several months in one go if you like, then check the import history for anything that failed.
Before you close the file
- Opening plus credits minus debits equals closing.
- No repeated header rows remain.
- Every date sorts correctly.
- Credits and debits are clearly separated.
- The sheet is named by account and month, with the original saved beside it.
Useful ideas start with a clearer picture of your own spending.
Explore your reports

