Reconcile two spreadsheets without losing two days to it

It is month-end. The source file says one thing. Finance's spreadsheet says another.
Somewhere in hundreds of thousands of rows there are a handful of breaks, and you are the one who has to find them.

Send me the two files. I send back a break report. The first one is free.

Get a free break report

Why month-end reconciliation still eats days

Nobody plans to reconcile by hand. It just happens.

The file is too big for Excel, so it gets split. Then the split gets checked.
The columns have different names on each side, so someone builds a lookup. Again.
One side rounds to 2 decimals, the other to 6. Every row "breaks".
VLOOKUP matches the first hit, not the right one. Duplicates hide.

The cost is not the two days. It is the break you missed, signed off, and found in next month's numbers.

How to reconcile two files properly

Match on a composite keyFund + ISIN + date, or account + security + as-of date. Not row order. Not the first VLOOKUP hit.

Map column A to column B onceTheir "MktVal_GBP" is your "Market Value". Written once, reused every month.

Compare with toleranceNumbers, dates and text compared by type, with the tolerance you set. Rounding noise stops burying real breaks.

Get a break reportMissing rows, extra rows, duplicate keys and every value that moved, both sides shown. Something you can hand to audit.

Compare CSV, Excel, XML, JSON and PDF

PDFXLSXCSVXMLJSONCSV inside JSON/XML

The source arrives as XML with a CSV buried inside it. Finance works in Excel. Someone forwards a PDF statement.
You should not have to convert any of it by hand first. Large-file processing is scoped after reviewing size, format and environment.

What you get back

A list of every break, grouped by type: missing, extra, duplicate, value difference.
Both values side by side, so you can see the difference without opening two files.
The rules that produced it, so next month's run is the same check, not a new one.

Keep the agreed rules so the check can be repeated and reviewed.

How the free proof works

This is a service-led first comparison: two files, one break report, no charge. Further work is agreed and quoted separately. Confidential files can be processed in a client-controlled environment by agreement. Do not send files or credentials through this enquiry form.

Tell me what you reconcile and how often (form below).
Agree the format, scope and safe processing or transfer method first. Anonymised or sample files are fine.
We agree the key and the column mapping. I run it and send the break report.

If it finds no breaks, the tested inputs agree under the agreed rules and scope. This does not prove all data is correct.

Questions about reconciling two spreadsheets

How do I reconcile two spreadsheets with different column names?

You map them once. Their "MktVal_GBP" becomes your "Market Value". The rule is saved and reused every month, so nobody rebuilds it at month-end.

Can it match on more than one column?

Yes. Fund + ISIN + date, account + security + as-of date, whatever identifies a row for you. Matching is on the key, never on row order.

What file formats can I compare?

PDF, XLSX, CSV, XML and JSON, including a CSV buried inside XML or JSON. The two files do not need to be in the same format.

How big can the files be?

Large files can be scoped. Confirm size, format, processing environment and turnaround before sending files.

Is the free proof really free?

Yes. Two files, one break report, no charge. Anonymised or sample data is fine.

What if it finds no breaks?

You have evidence of agreement for the tested inputs, scope and rules, rather than a guarantee that all data is correct.

Get your free break report

Useful next reading

See a synthetic break report example and how column mapping works.