CSV columns shifted in Excel: how to find and fix misaligned rows safely

Instead of searching row by row, use Repair CSV problems to see each problem, its original data, the suggested fix, and the reason. The input must be a CSV file, not an Excel workbook.

1. 1Choose a CSV: export Excel data as CSV first, then select that file.
2. 2Check for problems: find blank rows, inconsistent column counts, and exact duplicate candidates.
3. 3Review fixes: compare the original and proposed fix with nearby rows, then select what to apply.
4. 4Save: review selected and unresolved counts, then create a CSV separate from the original.

Misaligned CSV columns usually come from delimiters, quoting, row widths, or Excel import settings. Start from what you see on screen and inspect the corresponding part of the CSV.

For a small file, import it through Excel and confirm the delimiter and encoding in the preview. For a large file, use the check results to narrow down affected rows instead of scanning the whole file manually.

A trailing comma can mean that the last field is intentionally empty. Both lines below contain three columns.

Product,Amount,Notes
Product A,100,

Repair CSV problems uses the first row as the header, compares every row's column count with it, and shows why a row was flagged. It does not guess from the most common row width or from the values themselves.

It finds completely blank rows, rows longer or shorter than the header, clearly extra empty cells, and duplicate rows whose column count and every parsed cell value match exactly. You can compare the original with the proposed fix and view up to two surrounding rows on each side.

Each CSV can be up to 50 MiB. Major encodings such as UTF-8 and Shift_JIS are supported, along with comma, semicolon, tab, and pipe delimiters.

• Product A,100, has the same three columns as the header, so it is not a problem.
• Product B,200,Note,Extra value has four columns and is a review item that starts unselected.
• The completely blank row is a suggested fix selected by default, and you can clear it.
• The second Product C,300, is an exact duplicate candidate and remains unchanged until you select it.

Product,Amount,Notes
Product A,100,
Product B,200,Note,Extra value

Product C,300,
Product C,300,

A completely blank row starts selected. A row also starts selected when every cell beyond the first row's column count is empty. Short rows, extra values, and exact duplicates do not.

The output is a UTF-8 CSV with a BOM, which Excel commonly recognizes. Excel may still convert values on opening, so check the following before importing the file elsewhere.

• Confirm that every row aligns with the header.
• Check that leading zeros, dates, phone numbers, and long identifiers were not auto-converted.
• Test a small batch before importing into a business system.
• Keep the original file until verification is complete.

LocaCSV checks column counts, blank rows, and identical rows. It does not decide whether the values are correct or whether the first row contains the right column names.

• The values are correct.
• The CSV follows the destination system's rules.
• Each suggested duplicate is safe to remove.

• Convert encoding & format →Use this when garbled text or UTF-8 BOM is the main issue.
• Remove duplicates →Use a dedicated key-column rule instead of exact full-row matching.
• Filter rows →Keep only rows that match explicit conditions.
• Split by rows or size →Divide a CSV by row count or target size.
• Merge vertically →Stack multiple CSV files with the same columns.