How to use CSV Diff and Change Report
- Paste the earlier and newer tables, or choose a CSV, TSV or XLSX file for each side. Select workbook sheets where needed.
- Choose unique-key matching or row-position matching. For keys, identify each side's key column and decide how spaces, case and duplicates are handled.
- Choose the before/after or compact preview. Optionally name shared numeric columns for exact decimal totals.
- Compare the versions, review counts and changed cells, then download csv-changes.csv with the selected formula-prefix policy.
Example: CSV Diff and Change Report
Find one added row, one removed row and one changed name by the id column.
Options
- Before CSV text
- The earlier version of the table. Paste it here, or leave this empty and choose a file.
- After CSV text
- The newer version. A key that appears only here is reported as added.
- Match rows by
- Position compares the first row with the first row. Insertions shift later rows; no content-based realignment.
- Columns to total (optional)
- Comma-separated headings present on both sides. Exact decimal totals; blank or nonnumeric cells are refused.
- Before XLSX worksheet
- Exact name or one-based number; blank uses first visible sheet.
- After XLSX worksheet
- Exact name or one-based number; blank uses first visible sheet.
- Key column in the before table
- The column that identifies one row, such as an order number or email address. Enter it exactly as it appears.
- Key column in the after table
- The matching column name on the newer side.
- Compare cell values
- Comparison changes only what counts as a difference; the report still shows the real values.
- When a key repeats
- A repeated key cannot be matched one to one, so the default stops and names it instead of guessing.
- Protect spreadsheet formulas in the report
- A precaution for opening the report in a spreadsheet. A cell that begins with = + - or @, including after leading spaces, receives a prefix so it is less likely to run. The prefix changes the stored value, and an app can still treat the cell as a formula after a save and reopen, so it is not a guarantee.
Supported inputs and limits
- Each side: 20 MiB, 100,000 rows, 500 columns and 2,000,000 cells; change report: 50,000 rows.
- CSV/TSV or a selected XLSX sheet replaces pasted input.
- Repeated keys refuse one-to-one comparison unless first-row mode is selected.
- Row-position mode does not realign inserted rows.
- Decimal totals accept up to 100 characters and exponents from -30 to 30; blank or nonnumeric values refuse totals.
More details
The preview shows up to 60 changes; the guarded CSV holds all reported changes.
Reconciliation and cross-file matching remain separate tools.
Where your input is processed
This tool does not upload your inputs. See the limits for anything saved on your device.
A key turns a row comparison into a value comparison
Position mode pairs the first data row with the first, the second with the second and so on. Inserting a row near the top can therefore make later records look changed. A unique key such as an order ID lets this tool match the same record after it moves and report differences in its shared columns. Repeated keys need an explicit policy.
One row per changed cell
The download holds one row for each changed cell rather than one row per record, with the key, status, column name and both values. That shape is what makes a spreadsheet filter useful: sort by column to see every price change across the file, or filter to one key to review a single record's history.
Questions about CSV Diff and Change Report
Why does an unchanged row not appear in the report?
The report lists what changed. Unchanged rows are counted in the summary so the totals add up, and leaving them out keeps the file short enough to read.
Why is my renamed column treated as a removal and an addition?
The comparison matches columns by name. A renamed column is not found on the other side, so it is reported in the notes as present on one side only, and the remaining shared columns are still compared.
What does the status column hold?
added for a key only in the newer table, removed for a key only in the earlier table, and changed when a key appears on both sides but a shared value differs. An unchanged row is counted but never listed.
Can I compare files that use different delimiters?
Not in one run. The single delimiter choice applies to both sides, so set it to the delimiter that both files use, or convert one file first.
A key repeats in both files. Why was the comparison refused?
Rows cannot be paired one to one when a key repeats, so the default stops and names the key. Choose the first-row setting to compare the first row of each key; the summary then states how many rows were left out.