Usage

const fs = require ('fs')
const zlib = require ('zlib')

const {TarCsvReader} = require ('untar-csv')

const reader = TarCsvReader ({

//  test: entry => entry.name.endsWith ('.csv'),
//  delimiter: ',',
//  skip: 0,             // header lines
//  fileNumField: '#',   // how to name the file # property
//  rowNumField: '##',   // how to name the line # property
//  empty: null,

//  recordClass: Array   // to get [id, name] instead of {id, name}
//  columnClass: class extends CSVColumn {...}

    columns: ['id', 'name'],

//  header: 2,           // to read `columns` from first two CSV lines 
//  mask: 0b101,         // only the 1st and 3rd column will be read
//  maxColumns: 16384,   // leak protection: 16384 is the Excel limit

})

fs.createReadStream ('lots-of-data.tar.gz').pipe (zlib.createGunzip ()).pipe (reader)

for await (const {id, name} of reader) {
// do something with `id` and `name` 
}

Options

Most options are effectively passed to CSVReader, see there for details.

|Name|Default value|Description| |-|-|-| |columnClass|CSVColumn|| |columns| |Array of explicit column definitions (mutually exclusive with the header option)| |delimiter|','|Column delimiter| |empty|null|The value corresponding to zero length cell content| |fileNumField | null | The name of the file # property (null for no numbering) |header|0|Number of lines to gather columns definition| |mask|0|Unless 0 with header set, only masked columns will be read| |maxColumns|16384|Maximum # of columns in the header (to prevent a memory overflow)| |recordClass|Object|May be set as Array| |rowNumField | null | The name of the line # property (null for no numbering) |skip|0|Number of top lines to ignore (prior to header)| |test| ({name}) => true |tar entry filter, structure described at tar-stream|