Getting started

Expense manager with CSV import: columns, dates and a safe test run

How CSV import works in an expense manager: the columns it needs, date and amount traps, and a three-row test before you load months of data.

Editorial illustration for Expense manager with CSV import: columns, dates and a safe test run

CSV import is the most dependable way to move old records into an expense manager, because a CSV is plain rows and columns with nothing hidden. The catch is that three things go wrong quietly: dates that read the wrong way round, amounts with commas or currency symbols, and debits and credits sharing one column.

What does a CSV need to look like?

Most tools want one row per transaction with a date, a description, an amount and ideally an account. Open your file in Excel or Google Sheets and check it looks like this.

DateDescriptionAmountType
05-04-2026Grocery store1,240.00Debit
06-04-2026Salary credit55000Credit
09-04-2026Electricity bill2310.50Debit

Notice the problems in that sample: the first amount has a thousands comma, and the type column is separate from the amount. Those are exactly the things to tidy before upload.

What are the common CSV traps?

  • Dates: 05-04-2026 is 5 April in India but 4 May in some software. Confirm which your tool expects using one row where the day is above 12, such as 13-04-2026.
  • Amounts stored as text: if you cannot add the column with a sum, the numbers are text. Remove commas and symbols with find and replace.
  • Debits and credits: either a signed amount or two columns. Match what the tool asks.
  • Blank rows and totals: delete the grand-total line at the bottom or it is imported as a transaction.
  • Opening balance rows: remove them, they are not spending.

How do you run a safe test import?

  1. Save a copy of your file; never edit the original.
  2. Cut it down to three rows: one debit, one credit, one with a day above 12.
  3. Import those three and look at the dates, signs and totals.
  4. If all three are right, import the rest in monthly chunks rather than one giant file.
  5. Compare each month's total with your source before moving on.

What can the arithmetic tell you?

Suppose your old sheet's January column adds to ₹47,360 in debits. After import, the tool shows ₹46,120. The gap is ₹47,360 - ₹46,120 = ₹1,240, which matches the first row in the sample above, so the grocery line with a thousands comma was read badly or skipped. Fix the format, delete the month and re-import. The reconciliation takes two minutes and catches what no preview will.

Where does SpendixAI come in?

SpendixAI accepts statements as PDF, CSV or Excel. You pick the account before importing and review rows before they are saved, with possible duplicate imports flagged and an import history with status. Do not assume every layout reads perfectly; the three-row test above is how you find out. You can also download CSV reports later, so your data is never stuck.

What should you do with columns that do not fit?

Many old sheets carry extras: a notes column, a payment mode, a running balance. If the importer does not know a column, it either ignores it or gets confused, so decide what each extra is worth. Keep the notes by pasting them into the description column after the merchant name. Drop the running balance, since it is derived data. Keep the payment mode only if it maps to an account label such as cash or card. A tidy file with four columns imports better than a rich file with twelve, and you can add detail back later by editing individual transactions.

Pre-import checklist

  • Dates in one consistent format and tested with a day above 12.
  • Amounts numeric, with no symbols.
  • No totals, balances or blank rows left.
  • One account per file where possible.
  • Base currency chosen if you have more than rupees.
SPENDIXAI JOURNAL

Useful ideas start with a clearer picture of your own spending.

Explore your reports