Skip to content

Incorrect parse with trim: true and utf16le encoding #349

@Faris-Nasybulin

Description

@Faris-Nasybulin

Describe the bug

When using the trim: true option with utf16le encoding, the content is modified.

csv@5.5.3 contain csv-parse@4.16.3
Also reproduced for csv@6.1.3, csv-parse@5.2.0.

To Reproduce
In csv-parse@5.2.0:

const assert = require('assert')

const {parse: csvParse} = require('csv-parse/sync')

const data = 'ф, ы'
const encoding = 'utf16le'
const delimiter = ','
const record_delimiter = '\n'

const records = csvParse(Buffer.from(data, encoding), {
  //delimiter: Buffer.from(delimiter, encoding),
  //record_delimiter: Buffer.from(record_delimiter, encoding)
  encoding,
  trim: true
})

const correctRecords = data
  .split(record_delimiter)
  .map(
    record => record
      .split(delimiter)
      .map(str => str.trim())
  )
assert.deepStrictEqual(records, correctRecords)

Additional context

I need to parse utf16le encoded csv file with custom delimiter and record_delimiter. Please release bugfix also for csv-parse@4.16.3.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions