Compare Two CSV or XLSX Files
Compare two spreadsheet files using a selected key and group records into matched, missing, duplicate and conflicting buckets.
Mapping
Map Columns with Different Header Names
Suggest column mappings between two files using normalized header names, business synonyms and sampled data types.
Processing boundary: text and files are handled only for this request. Original files are not overwritten. Generated CSV downloads use expiring private tokens.
What this tool does
Two files can describe the same records while using completely different column names. The Smart Schema Mapper compares normalized header text, a curated set of common business synonyms and the detected type of sampled values. It then proposes the most likely Source B column for each Source A field and assigns a confidence score. Suggestions such as Emp ID to Employee Code or Inv_No to Invoice Number reduce manual setup, but every mapping remains a recommendation that a human must approve before reconciliation.
How to use it
- Upload the two datasets whose columns need to be aligned.
- Review each suggested pair, confidence score and explanation.
- Record or correct the mappings before running a comparison.
Limitations and review points
Delivery 1 uses deterministic rules rather than generative AI. Ambiguous fields, sparse columns and organisation-specific abbreviations may need manual correction.
Frequently asked questions
Does it automatically change my files?
No. It only displays mapping suggestions.
How is confidence calculated?
The score combines normalized header similarity, known business synonyms and detected value type.
Can two Source A columns map to the same Source B column?
The suggestion engine may show that in ambiguous data; review and correct it before use.
Related spreadsheet tools
Key Column Finder
Score spreadsheet columns by completeness and uniqueness to identify likely employee, invoice, asset or transaction keys.
Spreadsheet Column Profiler
Profile column types, blanks, uniqueness, maximum text length and sample values before mapping or importing a dataset.