Invalid row index doesn't reflect original row count

[jamescoletti]

This doesn't seem very intuitive at first glance. If it's by design, perhaps the docs should be updated.

When an invalid row is encountered, it appears to be removed from the original internal row count before the data-invalid event is emitted. This means each index value really corresponds to the total number of valid rows, not total rows, which presumably the user would need to keep track of while parsing.

Thanks for your work on this library.

var csv = require('fast-csv');

var csvData = "id,name,location\n\
1,Joe,Boston\n\
2,Emily,San Francisco\n\
,Harold,Dallas\n\
,Kate,Seattle\n\
5,William,New York";

csv.fromString(csvData, {headers: true})
  .validate(function (data) {
    return data.id;
  })
  .on('data', function (data) {
    console.log(data);
  })
  .on('data-invalid', function (data, index) {
    console.log(data, 'Error line: ' + index);
  });

/**
Outputs:

{ id: '1', name: 'Joe', location: 'Boston' }
{ id: '2', name: 'Emily', location: 'San Francisco' }
{ id: '', name: 'Harold', location: 'Dallas' } 'Error line: 2'
{ id: '', name: 'Kate', location: 'Seattle' } 'Error line: 2'
{ id: '5', name: 'William', location: 'New York' }

**/

[doug-martin]

This will be fixed in v8.2.0 #268. Ill close this once the new version is published.