Text files end each line with invisible characters. Windows traditionally uses CRLF (\r\n); Unix and macOS prefer LF (\n). CSV is not special: mixed endings create duplicate-looking rows in some tools and noisy Git diffs even when the data is identical.
When it matters
- Strict parsers that treat CR as part of the last column value.
- Hashing or checksum pipelines where whitespace changes the fingerprint.
- Cross-platform teams where one side re-saves CSV in Notepad and the other in vim.
Normalize deliberately
- Configure
.gitattributesfor text exports if the repo should enforce LF. - Use editor or CLI tools to convert endings before handoff, not halfway through a review chain.
After normalization, open the file in Table and confirm row counts match expectations. Search for stray \r characters if a column still looks padded on one platform only.