CSV Tools FAQ

No. The CSV data, file name, headers, and cell values are processed only in your browser and are not sent outside it.

The Free version accepts one file up to 50 MiB. You can move through every row in batches of 100 and every column in batches of 50. Long cells are shortened only on screen.

It can auto-detect UTF-8 and Shift_JIS, or you can choose manually from seven supported encodings. Viewer warns you when automatic detection is uncertain.

It shows the delimiter, BOM, and rows whose column count differs from the header. If broken quotes prevent parsing, it stops with an error instead of guessing.

Use the next-task links to convert encoding, split the file, repair CSV problems, select columns, or remove duplicates. You reselect the file after moving to another tool.

It is the column used to match the same row in both files. Choose a shared value that uniquely identifies each row, such as an ID, SKU, or email address.

Comparison stops because the row cannot be matched reliably. Fill or remove the blank key, then compare the files again.

Comparison stops because the rows cannot be matched one to one. Review or remove duplicate keys first.

No. Both files must have the same column names in the same order. LocaCSV stops rather than guessing how different columns correspond.

A key found only in the after file is added; one found only in the before file is removed. A matched row is changed when any non-key cell differs, otherwise it is unchanged.

No. Rows are matched by the selected key, so order alone does not create a change when the matched cell values are the same.

Comparison cannot run when the files exceed 50 MiB combined, cannot be decoded or parsed, have different headers, or contain blank or duplicate keys.

No. Both files, their names, the selected key, and all differences are processed only in your browser.

It detects rows with a different column count from the header, completely blank lines, exact duplicate rows, broken quotes, and other structural CSV errors.

It offers safe candidates that do not require guessing the data, such as removing completely blank lines or extra empty cells at the end of a row.

LocaCSV shows the cause and location, then stops without creating a partial output. You can check the input settings or source data and try again.

No. If a repair would require guessing where a quote should close or what a value should be, LocaCSV leaves it for review instead of changing it.

No. Only the repairs you select are applied. You can review the affected rows and counts first, and unselected rows remain unchanged.

No. The result is saved as a separate CSV, so the original file remains unchanged on your device.

No. The setting controls data rows per file; the header is repeated separately.

No. New CSV files and a ZIP are created in your browser.

No separate method is needed. The menu shows both the column position and header name.

Put them in a separate CSV, exclude them, stop, or replace them with a specified value.

Only the first file's header is written; later files add data rows.

Yes. Each file is detected and converted to the output encoding you choose.

The left row remains and the unmatched right-side fields are blank.

A key join adds columns; a vertical merge appends rows.

Yes. Choose UTF-8 with or without a BOM; the BOM option is convenient for Excel.

Encoding conversion preserves values. Typed JSON columns change type; blanks remain empty strings.

Yes. You can nest up to 100 conditions in eight levels of AND/OR groups.

Use the case-sensitive option. When off, uppercase and lowercase match.

No. Header text is preserved from the source CSV.

No. A new CSV containing the selected columns is created.

The first occurrence is kept; later rows with the same key are excluded.

Yes. Select multiple columns for the duplicate key.

Yes, when text order matches date order, as with ISO dates.

Yes. Move keys up or down; the top key is compared first.

Columns used for Sum, Average, Minimum, or Maximum must contain numbers such as “1200” or “-12.5”. Remove commas, units, and spaces around the number, as in “1,200 yen”. Blank cells can stay blank. To count text, choose “Row count” or “Non-empty count”. The displayed “data row ○, column ○” shows where to make the correction. Sums and averages can have rounding differences, so check exact totals separately.

Row count includes all rows; non-empty count includes rows with a value in the selected column.

No. Keep the source CSV if you need the detail rows.