Split a CSV by Data Records and Repeat the Header in Every File

Complete the task

1. Read a headered CSV or TSV with the correct delimiter. Review the parsed data-row count rather than relying on the number of physical text lines.
2. Set Data rows per file to a whole number from 1 to 20,000. The header is additional and does not count toward that value; choose a size that produces no more than 100 files.
3. Split into files, inspect the total and download each numbered part. Review the export protection and optional UTF-8 BOM before downloading. Parts are comma-delimited CSV even when the input uses another separator.

Worked example

The built-in order table has 5 data records. Keep the default size of 2 and split to produce part-001.csv, part-002.csv and part-003.csv with 2, 2 and 1 data records respectively. Each file also contains the four-column header. The repeated 00101 order appears in part-002 because splitting preserves source order and does not deduplicate. A line break inside a quoted field belongs to that record and does not consume another record slot.

Method and limits

This is a record-count split, not a byte-size, group-by-column or Excel worksheet split. The source must have a consistent header and row width and fit within 5 MiB, 20,000 data records, 200 columns and 250,000 cells including headers. At most 100 parts can be created; increase the rows per file if necessary. A header-only table produces no parts. Download each part separately; no ZIP archive is generated. Optional export protection starts enabled and can prefix formula-like values, including negative numbers, with an apostrophe. CSV output cannot enforce Excel cell types, and delimiters or line endings can differ from the source.