CSV Splitter — Split by Row Count or File Size

About this tool

Split large CSV files to meet the row-count or size limits of the system you are importing into. Multibyte characters, line breaks inside quoted cells and quotation marks are preserved. Size limits are enforced using the actual encoded output bytes. The entire input is validated before anything is saved. All processing happens locally in your browser; no data is uploaded.

Input file and format

Select a file to begin.

Input format will appear after you select a file.

Input preview (check encoding)

      

Only the start of the file is shown. Long records may be truncated in the preview, but the full records are processed when splitting.

Split limits and output format

Split and save settings

The limit includes each file's header, BOM and line breaks. 1 MB = 1,000,000 bytes.

Uncheck this option if your CSV has no header.When unchecked, the first record is also treated as data. When checked, the header does not count toward the selected data-row limit.

Line breaks, quotes, spaces and leading zeroes inside cells are preserved. Conversion stops with an error if the selected encoding cannot represent a character exactly.

Choose a file, review its preview and set the split limits.

How to use

  1. Choose a CSV and check the input preview for garbled text or wrong delimiters. Without a BOM, encoding detection is an estimate, so specify the encoding if needed.
  2. Set the maximum file size or the maximum number of data rows, excluding the header, to match your import limits. If the destination counts the header toward its row limit, subtract one from its allowed row count.
  3. Choose whether to repeat the header, plus the output encoding and line endings. Select “Validate and split”.
  4. When saving to a folder, the tool creates a dedicated subfolder and names output files sequentially, such as original-name_part_000001.csv. For individual downloads, confirm each completed download before proceeding.

Split behavior and limitations

  • Size is measured on the uncompressed CSV data, using the actual encoded bytes for multibyte characters. If a header and one data record would exceed the limit, splitting stops before saving and reports the required size.
  • Line breaks inside a quoted field belong to the same record. Line-ending settings affect record separators only, not cell contents. “Same as input” preserves mixed record endings and the presence or absence of a final newline.
  • Blank records and cells are not removed, and no conversion of numbers, dates or formulas takes place. Column names and field counts are not modified. Verify any column-format requirements of your import target separately.
  • The input is read twice: the first pass validates syntax, encoding, representability and split sizes; the second pass writes the output. No output files are created during validation.
  • For very large files, prefer direct folder saving in a compatible desktop browser such as Chrome or Edge. The entire input is not loaded into memory. Ensure sufficient disk space and keep the browser tab open until the task finishes.
  • Individual downloads are limited to 128 MiB per output file. If this is insufficient, reduce the size or row limit or use folder saving in a supported browser. A single record may be up to 64 MiB to protect the browser from excessive memory use.
  • If saving fails or is cancelled, the tool attempts to delete the dedicated output folder it created. If cleanup fails, the folder location is shown. Already-downloaded files cannot be removed automatically; do not import any incomplete output.

Convert between CSV / TSV and JSON