CSV Row Chunking

RowChunking splits text into one chunk per physical line.

RowChunking creates one chunk per physical line. It does not parse CSV quoting: a quoted field containing a newline is split across chunks. Parse or normalize multiline records first, or supply a custom chunker.

With clean_rows=True, whitespace is normalized even inside quoted fields. With False, leading and trailing whitespace is still stripped.

ParameterTypeDefaultDescription
skip_headerboolFalseSkip the first row before chunking. Row numbering in chunk metadata still starts from the original file's row positions.
clean_rowsboolTrueNormalize internal whitespace in each row. When False, rows are only stripped of leading and trailing whitespace.