Compare Excel Files

Compare Excel Files reviews changes between an original XLSX workbook and an updated XLSX workbook. It separates worksheet pairing from record matching, because a workbook can contain several unrelated tables with different keys. The result identifies added, removed, modified, and unchanged records within the confirmed scope, together with column-structure changes and sheets that were not paired.

Before importing, decide what a record represents and which fields identify it. An invoice number may be sufficient for an invoice table but insufficient for invoice lines. A line table might need invoice number plus line number. Comparing by a business key keeps ordinary row reordering from appearing as a change, while comparing by position intentionally answers a different question about corresponding row locations.

Select exactly two workbooks. The first is shown as Original version and the second as Updated version. Reversing these roles reverses the meaning of additions and removals. The tool does not infer which file is newer from its name, modification time, or a date appearing in a worksheet. Confirm the version roles before interpreting the report.

Each workbook lists its worksheets and visible state. Select only the tables relevant to this comparison and confirm their ranges, header rows, data starts, and hidden-content settings. A visible sheet can still contain hidden rows and columns, so worksheet visibility alone is not enough to define scope. Unselected content remains outside the comparison, even when the result processes every record within your selected regions.

Suggest same-name pairs uses equal worksheet names as candidates. You can add a manual pair when a tab was renamed or select a different updated sheet for an existing pair. The pairing does not imply that the schemas or keys are correct. Review each pair and configure its matching fields independently rather than applying one key to the entire workbook.

A sheet can belong to only one pair. Old-only and new-only sheet names are reported separately as removed or added when there is no counterpart. Sheets that exist on both sides but were not selected for a pair are labelled unmatched. These distinctions keep a deliberately excluded sheet from being presented as a deleted business table merely because it did not participate in the selected comparison.

Range selects the worksheet area to consider. Header row and Data start row establish column names and the first business record. They can differ between versions. For example, an updated export may insert an introductory line above the header without changing its business records. Correctly selecting each header prevents that new introduction from creating false column or row changes.

Changes to worksheet settings invalidate the previous result and require confirmation again. The active pair must follow those updated selections rather than retaining a hidden copy of the old range. Use the Comparison scope report to verify the actual range, header, data start, hidden-row treatment, hidden-column treatment, and value mode used on each side. That report also travels with an all-results export.

Original keys and Updated keys allow composite keys in selection order. Select the corresponding fields on both sides in the same semantic order. The comparison uses unambiguous key representations, so values that happen to contain punctuation cannot create collisions through simple string concatenation. It does not silently trim spaces or normalize identifier spelling for this task.

Duplicate or empty keys block deterministic matching. Choosing an arbitrary first occurrence could pair the wrong customer, transaction, or product and produce misleading differences. Resolve the source identity issue or deliberately use row-position comparison if positions are the intended identity. Position comparison is an explicit checkbox; it is not an automatic fallback that disguises an ambiguous key problem.

Stored values / saved results compares the imported underlying representations and saved formula results. Displayed text compares the formatted text available in the workbook. These modes can legitimately disagree: one cell may store 123 and display 00123, while another stores 123 and displays 123. Select the mode that answers your review question and retain that mode in the report. Stored value comparison also distinguishes numbers, Boolean values, and text: numeric 1 and text 1 differ, and a numeric key does not match a text key. The Stored type changes report identifies those changes while keeping the old and new values readable. Date-formatted numeric cells remain stored serials in this mode; compare displayed text if the calendar representation is the intended evidence, and check each workbook’s date system.

Formula text compares expressions as text without calculating them. Two expressions may produce the same saved result and still differ in this mode. Conversely, matching formula text does not prove that the resulting business values are current. Relative references are not normalized into mathematical equivalents, and rearranged expressions are not declared equal through algebra. Non-formula cells still contribute their imported values.

A formula without a usable saved result is not an ordinary empty cell. In value-based comparison, the affected pair is reported as indeterminate rather than allowing two missing caches to count as equal blanks. Formula-text comparison can inspect an available expression even when its cached calculation is missing or contains an error, because the expression itself is the selected evidence.

The tool does not update external links, run macros, call connections, or calculate formulas. A workbook saved before its latest calculation can contain stale caches. If the review requires current calculated values, use the source application to calculate and save a new copy, then repeat the comparison. Keep the distinction between imported saved results and a fresh calculation visible when explaining the outcome to another reviewer.

The original Data sheet contains id 00123 with amount 10 and id 00456 with amount 20. The updated version lists 00456 first and changes 00123 to amount 11. Pair the sheets, select id on both sides, and compare stored values. The expected result is one modified record and one unchanged record, with no addition or removal caused by the row order change.

The modified-cell report retains amount, old value 10, new value 11, and both source positions. If the same cells instead contain two different formulas with identical saved results, value mode may report the read results as unchanged while formula mode identifies the expression difference. That is not contradictory: the two runs ask different questions and their mode labels preserve the distinction.

The main result summarizes each selected pair. Use Result scope to open a pair's added records, removed records, modified cells, unchanged records, or column-structure report. Each report keeps its relevant source locations. The unmatched-sheet view explains sheets outside the pair list, and the scope view explains the exact selected regions rather than claiming every workbook object was compared.

Compare columns can restrict the fields under review. Leaving that selection empty compares common fields. New or removed columns are reported as structural differences separately from changed cell values. A zero-change result is valid when the selected fields and mode match. Check scope before concluding that the workbooks are interchangeable: styles, colors, drawings, charts, macros, protection, and other non-data features are not part of this comparison.

Choose a specific report in Export scope or select All results and reports ZIP. A current report, all paired sheets, and an individual change status are different scopes; the grid's search box does not silently switch between them. CSV uses UTF-8 and an optional BOM. XLSX output is a new data workbook, with selected representations stored as text cells rather than copied live formulas.

Verify export and show summary reads the output back. Formula-like CSV text can receive explicit apostrophe protection, with the affected count shown before downloading. Changing inputs or rules makes the result stale and disables export until it is rerun. Cancel stops the worker, and bounded package, range, cell, and time limits prevent unbounded jobs. Language changes preserve the selected keys and data interpretation; refreshing or closing clears the local session.