tiddlycsv is a small, streaming CSV parser. It's about ~1.3kb, has zero dependencies, and is tree-shakable.

Doesn't care about headers, keyed rows, anything but strings. Just supports the CSV spec including multi-line and quoted strings. Supports \r\n as well as \n for line separators only (i.e., won't rewrite \r\n to \n within multi-line strings).

Usage

Install tiddlycsv. Includes types.

Methods

• parseCSV: parses a whole string CSV into an array of arrays

• iterCSV: returns a Generator for each line of a whole string CSV

• streamCSV: returns a ReadableStream for each line of a streaming CSV

• streamCSVChunk: returns a ReadableStream for chunks of the streaming CSV (faster)

• buildCSVChunkStreamer: builds a low-level function which converts incoming CSV data into rows

• writeCSV: simple method that converts array data into CSV string output (call multiple times to "stream")

In general, if the read helpers don't work for you, use buildCSVChunkStreamer and provide data your way. Minifying that method only is about ~0.55k.

Speed

It's fast for its size. To parse multiple copies of 1.csv:

tiddlycsv: 837.074ms (baseline) ~1.3kb
udsv: 540.037ms (0.63x) ~5kb
but-csv: 737.836ms (0.88x) ~0.5kb
papaparse: 1.120s (1.34x) ~20kb

Output

Unlike other parsers, tiddlycsv will return empty strings for trailing commas and empty lines. It skips the last input line if it does not have any data. For the input data:

Name,Age
Sam,37
Jim,

Anne,56

You'll get results like:

[["Name", "Age"], ["Sam", "37"], ["Jim", ""], [""], ["Anne", "56"]]

Should this be configurable? Maybe!