CODEWITHSUNDEEP
javascriptnode.jscsvnode-csv-parse
John Horner
John Horner

Reputation: 282

How to parse CSV with an invalid first line in node-csv

Given this problem CSV file:

"This is a header which doesn't match the CSV"
id,name
1,Australia
2,Austria
3,Belgium

how would I parse it in node-csv, skipping that first line, and using the second line as the column headers?

What I've tried:

1)

import fs from 'node:fs';
import { parse } from 'csv-parse';
let rows = [];
fs.createReadStream('./downloaded/bad_data.csv')
         // columns option set
        .pipe(parse({ columns: true })) 
        .on("data", function (row) {
                rows.push(row);
                //console.log(row);

        }

        ).on('end', function () {
                console.table(rows);
        });

this results in a non-recoverable error because of the quote.

2)

     // columns not set, skipping errors
    .pipe(parse({ skip_records_with_error: true }))

This parses fine but without the headers, output:

┌─────────┬──────┬─────────────┐
│ (index) │  0   │      1      │
├─────────┼──────┼─────────────┤
│    0    │ 'id' │   'name'    │
│    1    │ '1'  │ 'Australia' │
│    2    │ '2'  │  'Austria'  │
│    3    │ '3'  │  'Belgium'  │
└─────────┴──────┴─────────────┘

  1.    // columns true, skipping errors true
  .pipe(parse({ columns: true, skip_records_with_error: true }))

This produces no output.

4)

    // columns true, skipping true, starting from line 2
    .pipe(parse({ columns: true, from_line: 2, skip_records_with_error: true }))

this parses fine but misses the first line of the data:

┌─────────┬─────┬───────────┐
│ (index) │ id  │   name    │
├─────────┼─────┼───────────┤
│    0    │ '2' │ 'Austria' │
│    1    │ '3' │ 'Belgium' │
└─────────┴─────┴───────────┘

    // columns, skipping, start from line 1
    .pipe(parse({ columns: true, from_line: 1, skip_records_with_error: true }))

This produces no output.

What am I missing and how can I get my columns from line 2 if line 1 contains an error?

Upvotes: 0

Views: 54

Answers (1)

Jack
Jack

Reputation: 120

I have tried your way and its working fine

 const fs = require('node:fs');
    const { parse } = require('csv-parse');
    let rows = [];
    fs.createReadStream('temp.csv')
      .pipe(parse({ columns: true, from_line: 2, skip_records_with_error: true }))
      .on("data", function (row) {
        rows.push(row);
      }
    
      ).on('end', function () {
        console.table(rows);
      });

Here's the output

┌─────────┬─────┬─────────────┐
│ (index) │ id  │    name     │
├─────────┼─────┼─────────────┤
│    0    │ '1' │ 'Australia' │
│    1    │ '2' │  'Austria'  │
│    2    │ '3' │  'Belgium'  │
└─────────┴─────┴─────────────┘

Upvotes: 0

Related Questions