PRACTICAL GUIDE
How to compare two CSV files when the rows don’t line up
A line-by-line diff assumes each record is in the same position. That breaks when an export is sorted differently or has new rows. A key-based comparison matches the same record first, then checks its fields.
1. Choose an ID that means the same thing in both systems
Use a stable customer ID, SKU or record identifier. A name or email can change. An ID such as 001 is text: converting it to a number would lose its leading zeros. ExportCheck keeps keys as text.
Baseline New export record_id,name,plan customer_id,display_name,tier 001,Ada,start 002,Lin,pro 002,Lin,pro 001,Ada,pro
Map record_id to customer_id, and plan to tier. The correct result is one changed key and one unchanged key, despite the reversed row order.
2. Fix duplicate and blank keys before trusting matches
Two rows sharing an ID are ambiguous. Choosing the first or last silently can hide a change. ExportCheck marks the entire key group as ambiguous and lists its data-record positions. A blank key is reported separately. If multiple rows per ID are intentional, the batch toolkit can match a composite key such as order ID plus line-item ID.
3. Compare only the fields you mean to check
Export timestamps often change on every run. Uncheck those fields. If a header changes between systems, map it explicitly. The report lists columns that were not checked so that a green result cannot be confused with a check of the entire schema.
4. Normalize deliberately
Ignoring spaces or letter case can be useful for display names, but can hide real changes in codes or free text. These settings are off by default. Numeric tolerance is absolute: a tolerance of 0.01 accepts a difference of at most one hundredth, including the boundary. Only fields marked numeric use it.
Use plain decimal values: 1234.50, not $1,234.50. Currency labels, scientific notation and blanks are flagged for review rather than coerced. A numeric field may lose the significance of leading zeros; keep identifiers as text.
5. Save evidence and repeat the same rules
Download HTML for review and CSV for further filtering. Both include all comparison groups, even if the web view is filtered. CSV output prefixes potentially formula-like cells with an apostrophe to reduce spreadsheet formula execution; that is intentional and can alter how such values display when re-imported.
A saved settings file contains headers and comparison rules, not input records. Keep exported reports according to your organization’s data-handling rules. This tool checks differences; it does not certify business or accounting correctness.