Table Reconciliation

Guide and worked example

Compare inventory exports, staff lists or order records using stable identifiers rather than row positions. A record can still match after moving to another row.

1. Export both tables as CSV or TSV with headers. Select encoding and delimiter before uploading A and B.
2. Choose identifier columns on both sides, optionally a composite key. Select the fields to compare separately.
3. Comparison is exact text by default. Ignore surrounding whitespace or case only when your business rules permit.
4. Compare, inspect exceptions and export an exceptions or full CSV. Duplicate identifier groups require manual review rather than forced pairing.

Worked example

Download the samples below. Use UTF-8, comma delimiter, id as the identifier, and name and quantity as comparison fields. Expect 001 matched, quantity for 002 changed from 5 to 7, 003 only in A, and 004 only in B. Leading zeros are preserved as text.

sample-a.csv

id,name,quantity
001,Pen,10
002,Book,5
003,Folder,2

sample-b.csv

id,name,quantity
001,Pen,10
002,Book,7
004,Clip,8

Limitations and checks

Each file supports up to 5 MiB, 10,000 records, 100 columns and 500,000 cells. Headers must be unique and nonempty. Formulas are not executed; amounts and dates are not normalized.

Why do similar-looking values differ?

1 and 1.00, different date formats, and surrounding whitespace are different text by default. Check meaning before cleaning source data.

Why are duplicate identifiers not paired row by row?

Row order cannot establish identity. The whole duplicate group is isolated to avoid misleading matches.

Why is text garbled or all data in one column?

Check encoding and delimiter, then reselect the file after changing settings. Do not reconcile garbled data.

Files are processed in your browser without uploading their contents. Download before closing or refreshing.