CSV Integrity LabIn-browser processing · No file uploads

Instructions · Updated 2026-09-16

Keep IDs as text
starting at import

Product IDs are identifiers, not quantities to add. Preserve 00123 and long IDs at the point where the original file enters Excel.

Check the result with synthetic CSV files first

These invented records are the same as the built-in tool example. The changes in the saved file were deliberately constructed, not recorded from a particular Excel version. This tutorial has not been tested in a specific Excel version.

Download both UTF-8 practice files, or save them yourself using the steps below. Compare the downloaded files directly in this tool; do not first open and resave them in Excel.

Download original.csv (original sample)

Download saved.csv (synthetic saved sample)

Copy each complete block below into a plain-text editor and save it as original.csv or saved.csv using UTF-8. Keep the header row and every character. Do not open or rewrite these practice files in Excel first.

The sample keeps its original Chinese headers: column 1 is product ID, column 2 is batch ID, and column 3 is quantity. Every language uses the same CSV data.

Original file: original.csv

商品編號,批次編號,數量
00123,12345678901234567,8
00456,98765432109876543,12

Synthetic saved file: saved.csv

商品編號,批次編號,數量
123,12345678901234500,8
456,9.87654321098765E+16,12

CSV checker

Select original.csv in the first file field and saved.csv in the second, choose UTF-8 and a comma delimiter, then run the check. You can also use the sample button to load the same records directly.

Expected differences: 4 cells

The row numbers below include the header row. Check each row, column and complete string, not just the total number of differences.

Both product IDs lose leading zeros. One batch ID has changed trailing digits; the other uses scientific notation. The quantity values remain: 8, 12. These differences establish only that the strings differ between the files, not which application caused the changes.

Compare the original file with itself

Select original.csv in both file fields and run the check with the same settings. Expected differences: 0 cells. This means the parsed field strings match. Leading zeros and long IDs may still be flagged for attention; that does not contradict a zero-difference result.

Zero differences do not prove that the file was never damaged or certify overall data quality. Different enclosing quotes, a BOM or line endings can still produce zero differences when the parsed strings match. A single file cannot establish earlier values.

Then import your own CSV

The steps below are based on Microsoft documentation. Excel menus and automatic conversion defaults vary by version. Preserve the original export and work on a copy.

  1. Keep the original CSV. Never overwrite your only export. Inspect the source here or in a local text editor first.
  2. Use Excel Data → From Text/CSV. Depending on the version, this may be under Get Data. Enter the transform/edit data workflow first.
  3. Set ID columns to text. Review Power Query's applied steps and replace or remove earlier automatic numeric conversion. Converting an already truncated number back to text is insufficient.
  4. Spot-check after loading. Compare one leading-zero ID and one ID longer than 15 digits with the complete original CSV strings. If exporting again, save a new file and compare it too.

Why aren't quotes always enough?

CSV quotes enclose fields containing delimiters or line breaks; they do not declare Excel data types. Therefore "00123" still requires correct import settings.

What if digits are already lost?

Formatting and text conversion cannot infer unknown source values. Obtain the original export again. Without a backup, leave unknown values unknown instead of guessing a repair.

For menu paths and version differences, follow Microsoft's official documentation. This site does not run or control your Excel.