How to use Duplicate Records Across Files
- Choose two UTF-8 CSV files and use their order controls to set file A followed by file B.
- Enter key-column names present in both files, then choose whether to trim surrounding spaces or ignore letter case.
- Optionally include original source rows side by side. Near-candidate review supports one key column and one code-point edit; treat those pairs as suggestions.
- Inspect exact-key pairs separately from near candidates, then download the CSV or JSON review. No source records are merged or deleted.
Example: Duplicate Records Across Files
Two CSV tables with key values 001, 002 and 002, 003 respectively.
Options
- Key columns
- Comma-separated column names present in both files. Names containing commas can be quoted, as in CSV.
- Also list near candidates for manual review
- One key column only. One Unicode code point insertion, deletion or substitution after the chosen normalization, not a full grapheme comparison. Keys up to 80 code points; 120,000 signatures, 20,000 candidate checks and 5,000 total pairs. No automatic merging.
- Show both source rows side by side
- Adds the original values from both files to the matching-pairs report. This does not merge records; export is limited to 8 MiB.
Supported inputs and limits
- Two UTF-8 CSV files, 2 MiB each, 10,000 combined rows and 50 columns per file.
- Up to 5,000 matching pairs.
- Near candidates support one key column, one Unicode code point edit and bounded signature/comparison counts; they are never merged.
- Optional full source values use an 8 MiB CSV bound.
- Preview shows 100 pairs while exports contain every accepted pair.
More details
Selected-key equality does not prove all source values agree.
Where your input is processed
This tool does not upload your inputs. See the limits for anything saved on your device.
A pair is evidence for review, not a merge instruction
Row numbers count the header as row 1. Two matching keys can have different names, addresses or other data, and repeated keys can produce several row pairs. Open both original rows and compare the fields that matter to your task. The exported match list leaves the source files unchanged; its formula guard also means a spreadsheet cell may display an added apostrophe.
Questions about Duplicate Records Across Files
Will matching records be merged?
No. The output identifies pairs for your review.
Is a near match a confirmed duplicate?
No. Similar text can describe different records; inspect the selected fields and original data.
Why are leading zeros preserved?
CSV key fields are compared as text so identifiers such as 001 keep their entered form.