Import a bank CSV into Quicken Classic for Windows: layout and duplicate checks

Keep all seven columns, including the empty ones

A bank export with Date, Description and Amount won't fit this layout as it stands. Quicken requires seven column positions in a fixed order. Four need a value in every transaction; the others can be empty, but their positions must remain. To get a baseline from your installed application, choose Create Sample File from the import menu before preparing real data.

The FI Payee field is for compatibility with Mint exports; Quicken says it ignores its contents. The Debit/Credit field also comes from Mint compatibility. Quicken recommends leaving it blank and using signed amounts instead. Don't delete either column because you have no use for it: that would shift the remaining fields into the wrong positions.

A header is optional, but it still needs to follow the documented order if you include one. After Account, you can add Tag, Memo, Chknum, in that order, leaving empty intervening positions when you use a later field. For a first file, the seven-column layout below avoids those extra decisions. These requirements come from the official required-file-format section.

Write month first, and check what each sign means

Quicken specifies mm/dd/yyyy; leading zeros on the month and day are optional. In this layout, 03/04/2026 means March 4, not April 3. You need to settle the source's date convention before changing it, because replacing separators in an ambiguous date won't tell you which month it belongs to.

Quicken describes positive amounts as increasing net worth and negative amounts as decreasing it; an unsigned amount is treated as positive. Our example is a checking account, so a deposit is positive and a purchase is negative. Your export might use a different convention, so compare one known deposit and one known withdrawal against the bank records. Don't silently apply this checking-account example to card debt.

For a simple working file, use plain amounts such as +1000.00 and -25.00. Quicken allows grouping commas in amounts, but any field containing a comma must be quoted: "1,000.00" is one field, whereas an unquoted comma separates fields. The same rule applies to a payee such as "Office Supplies, Inc.". Quicken also recommends preparing the file with a text editor or spreadsheet tool that supports UTF-8.

These date, amount and quoting rules are in Quicken's additional format notes. Its web page displays a typographic dash in the negative example. Our CSV uses the ordinary keyboard minus sign, not a copied typographic dash. We haven't verified sign parsing inside Quicken, so compare it with the application-generated sample and test in a separate disposable data file before using real records.

A three-row file you can inspect

This fictional USD checking account has a header and three transaction rows. It's not customer data or a Quicken-generated sample. To inspect it as a file, copy the block into a plain-text file saved as UTF-8 with a .csv extension. The account name is deliberately Example Checking; don't import these fictional transactions into your real account.

Date,Payee,FI Payee,Amount,Debit/Credit,Category,Account
03/02/2026,Client payment,,+1000.00,,,Example Checking
03/03/2026,"Office Supplies, Inc.",,-25.00,,,Example Checking
03/04/2026,Phone service,,-75.00,,,Example Checking

The doubled comma after each payee leaves FI Payee empty. The three commas after each amount leave Debit/Credit and Category empty before Account. The comma inside the quoted office-supplies payee is part of its name, so that row still has seven fields. Category is blank on purpose: the example demonstrates an import layout, not how you should classify these transactions.

The imported movement should be 1000.00 - 25.00 - 75.00 = 900.00. With an independently supplied opening balance of 500.00, the arithmetic closes at 1400.00. The opening balance is context, not a fourth transaction in this file. These are expected arithmetic results, not observed Quicken output.

If your real file has the right rows but different columns, you can prepare the documented layout with the Quicken CSV formatter. Formatting doesn't establish that an import will succeed, and it can't tell you which transactions are already in your Quicken register.

Decide which rows are new before importing

Quicken's CSV warning is unusually direct: unlike Web Connect QFX imports, CSV imports don't check for or match existing transactions. A fresh filename doesn't make old transactions new. Importing our whole example a second time would, under the documented behavior, add three more rows and another 900.00 of movement. With no other changes, the illustrative balance would become 2300.00 instead of 1400.00. That's a calculated consequence, not a live repeat-import test.

The comparison belongs at the account level: work on one account and currency at a time, leaving the original bank export untouched. Before importing, compare the proposed rows with what's already in that account's register, including manual entries and downloaded transactions. A date boundary helps narrow the search, but it doesn't prove that every row on one side is new.

1. In a separate review copy, mark each proposed row as already present, new or unresolved. The account, date, amount, payee and any bank reference support that comparison; the date and amount alone aren't enough, since two purchases with the same date and amount can both be real.
2. Only rows you've confirmed are missing belong in the export. Keep review notes outside the Quicken CSV so they don't add unsupported columns. Unresolved rows need to be set aside for investigation, with the batch recorded as incomplete so they aren't forgotten.
3. The next person needs to know what was added, so record the target account, filename, selected source rows, row count and signed total for the batch. Keep the exact imported file after checking the result rather than relying on a filename containing final.

In the fictional file, if Client payment and Office supplies are already recorded, only the Phone service row belongs in the next import. With a starting register balance of 1475.00, that one -75.00 row brings it to 1400.00. Don't import all three and plan to remove duplicates later.

Multiple bank downloads need their own overlap review before this register comparison. The guide to merging overlapping bank CSVs explains why matching dates alone are insufficient. A clean merge can still overlap with Quicken, so it does not replace the register check.

Import once, then compare the actual result

Before changing real records, you need a backup of your data file. Use a separate disposable file to learn the import behavior without changing those records. The sequence below follows Quicken's published import steps, not a walkthrough we ran in the application.

1. With the intended Quicken data file open, select File → File Import → Banking Transactions → From (.CSV) File.
2. Review the format requirements, select Choose .CSV File to Import..., choose your reviewed file and select OK.
3. If the CSV account name doesn't exist in the data file, Quicken documents choices to ignore those transactions, add them to an existing account, or create a new account. Read that prompt carefully: a bank label that differs from yours isn't a reason to create a second account.
4. The import summary gives you a result to compare with your planned row count; Quicken says it also lists any new accounts and tags. In the actual register, check the dates, payees, amount directions and account placement, then compare the selected rows and totals with the source records.

An unexpected result is a reason to stop, not to select the file again. If you aren't sure whether rows were added, inspect the register first. Correcting a filename or header and retrying does not undo transactions already imported.

Quicken also states that CSV import doesn't detect or create transfers, even when Category contains a bracketed account name such as [Savings]. Having both sides in the file doesn't mean a transfer relationship appeared, so review transfers separately in the register. A matching ending balance is an arithmetic check, not confirmation that categories and account relationships are right.

Use a bank export before extracting a PDF

A bank-supplied QFX file is a different workflow from this CSV route, and Quicken's CSV instructions explicitly distinguish their matching behavior. Before reshaping a large historical file, check whether the bank offers a native Quicken download for the dates you need. The CSV, QBO, QFX, OFX and QIF guide explains the formats. Changing an extension is not a format conversion, and a bank download still deserves a register review.

If only PDF statements remain, ask the bank about historical transaction exports first. For a small gap, manual transcription is also an option. AnyStatement can extract PDF statements into reviewable transaction data when those alternatives don't fit. You still need to compare the extracted rows with the PDF before preparing the Quicken layout and reviewing overlap; extraction doesn't remove those steps.

What we checked, and what we did not test

We opened Quicken's Import transactions from .CSV file page on October 17, 2026, including its HTML, which records April 14, 2026 as the last-modified date. We parsed the fictional CSV with Python, confirmed seven fields in the header and every row, and recomputed the totals and balance examples with Decimal arithmetic. We didn't run live Quicken, generate its native sample file, test this CSV in Quicken, or test Excel or Sheets. The sample is a documentation-based preparation aid, not a certified import file. Your next checks are the installed application's sample and a test in a separate disposable data file.