@openimport/csv

Zero-dependency CSV / TSV format plugin for OpenImport.

• Streaming RFC 4180 parser on Web Streams — same code in browsers, workers, Node ≥ 20, Bun and Deno
• Quoted fields with delimiters, line breaks and "" escapes; CRLF / LF / CR; ragged rows
• Encoding detection: BOM (UTF-8, UTF-16 LE/BE), strict UTF-8, fallback to windows-1252 (Excel on Windows)
• Delimiter sniffing for , ; \t |, quote-aware
• Cheap row-count estimate for large files

import { createImporter } from '@openimport/core'
import { csv } from '@openimport/csv'

const importer = createImporter({ schema, plugins: [csv()] })

Options

| Option | Default | Description | | ----------- | ------------ | ----------------------------------------------------------- | | delimiter | sniffed | Force , ; \t \| or any single character | | encoding | detected | Any WHATWG encoding label, e.g. iso-8859-15 | | quote | " | Quote character | | chunkSize | 1000 | Rows per yielded chunk | | extensions| — | Additional file extensions to accept |

Low-level API

import { createCsvParser, parseCsv, sniffDelimiter, detectEncoding } from '@openimport/csv'

parseCsv('a;b\n1;2') // [['a', 'b'], ['1', '2']]

const parser = createCsvParser({ delimiter: ',' })
parser.push('id,name\n1,"Ann') // [['id', 'name']]
parser.push('e"\n')             // [['1', 'Anne']]
parser.flush()                  // []

createCsvParser accepts text split at any position — inside quotes, between \r and \n, or between an escaped "" pair. Blank lines produce [], so row indices keep matching file lines.