Getting started

PocketSmith not working? Fix the CSV before blaming the import

PocketSmith not working with an imported file? Fix the usual CSV problems: date formats, text numbers, split debit and credit columns, and header rows.

Editorial illustration for PocketSmith not working? Fix the CSV before blaming the import

When an import into a budgeting tool such as PocketSmith fails or produces nonsense, the problem is often the file rather than the tool: dates in the wrong order, amounts stored as text, or a bank file with extra header lines. Clean the CSV in a spreadsheet and most of these problems vanish. we cannot see how any tool reads your file, so check its own help for the layout it expects.

Problem one: dates that flip or fail

A file may hold 03/04/2026 meaning 3 April, and a tool set to the other style reads 4 March. Check a date where the day is above 12, such as 25/04/2026, to see how the tool interprets it. If you need to convert a date stored as text in dd/mm/yyyy form, put this in a helper column:

  • =DATE(RIGHT(A2,4), MID(A2,4,2), LEFT(A2,2)) builds a real date from text like 25/04/2026.
  • Format the result column as a date, then copy and paste it back as values only.

Problem two: amounts stored as text

Numbers with thousands commas or a rupee symbol are often text. Left-aligned numbers are the giveaway. In a helper column use:

  • =VALUE(SUBSTITUTE(SUBSTITUTE(C2,",",""),"₹","")) strips commas and the rupee sign, then turns the text into a number.
  • Check the total with =SUM(D2:D500) against the statement.

Problem three: debit and credit in two columns

Many bank files give money out in one column and money in in another. Some tools want a single amount column where spending is negative. In Excel or Google Sheets, with debits in C and credits in D, use:

  • =D2-C2 gives a positive number for money in and a negative one for money out.
  • If blank cells cause an error, use =N(D2)-N(C2) instead.

Problem four: junk above or below the table

  1. Delete account details, title lines and blank rows above the header.
  2. Make sure there is exactly one header row.
  3. Remove totals, page numbers and 'continued' lines in the middle.
  4. Save as CSV, then open it in a plain text view to check that it looks like rows of values.
BeforeAfter
25/04/2026, 1,250.50, Dr2026-04-25, -1250.50
02/05/2026, 8,000.00, Cr2026-05-02, 8000.00

A check that nothing was lost

Say the original has 112 rows, and the cleaned file shows 109 rows after you removed 3 header and total lines. 112 minus 3 is 109, so the counts agree. If the numbers do not agree, you deleted something you did not mean to. Also compare totals: credits minus debits from the file should match the statement.

When the file itself is the problem

SpendixAI accepts statements as PDF, CSV or Excel and shows you the rows before saving, so you can fix a wrong date or amount at that point rather than hunting through a sheet. You pick the account first and can upload several files in a batch, with an import history showing each file's status and an Import failed notification if one does not go through. It cannot promise that every layout will be read perfectly, which is why the review step is there.

What about Excel files with merged cells?

Merged cells and coloured header bands look tidy but confuse import tools. Select the whole sheet, unmerge cells, delete decorative rows, and make sure each column holds one kind of value. If your statement came as an Excel file, save a plain copy with only the header row and the transactions.

Before you try again

  • Fix dates, then amounts, then the header, in that order.
  • Compare row counts and totals.
  • Import a small file first, say one week.
  • Keep the cleaned copy and the original side by side.
SPENDIXAI JOURNAL

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

Explore your reports