wizcorp / csvparser

Drag 'n drop CSV parser

CSVParser

Drag 'n drop CSV parser

##Usage

Instantiate CSVParser with a config, the config should be structured as follows:

var CSVParser = require('csvparser');

var myConfig = {
	rules: myRules,
	tests: myTests,
	target: myTarget, // optional
	options: myOptions // optional
}

var myParser = new CSVParser(myConfig);

###rules

CSVParser uses rules to test and parse all values in your CSV. For each field in your CSV you must define a rule to tell CSVParser how to handle it.

var myRules = {
	id: 'string',
	name: 'string',
	age: 'number',
	married: 'boolean',
	birthday: 'date',
	department: 'departmentId'
}

###tests

CSVParser has a few built-in tests:

• boolean
• date
• number
• string
• timeString

You are free to create your own. It will need a test and parse function.

var myTests = {
	departmentId:
		test: function (value) { return departmentData.hasOwnProperty(value); },
		parse: function (value) { return value; }
	}
}

###target

CSVParser will create a table to show the result of validation of the parsed data. Provide a DOM element so it can appear in your page.

var myTarget = document.getElementById('personnelTarget');

###options

CSVParser has a few options that you can set:

• allowNull: boolean, defaults to true - null values are acceptable in all fields.
• allowUndefined: boolean, defaults to true - undefined values are acceptable in all fields.
• empty: defaults to undefined - What you want to replace empty values with.
• optional: array of keys - Which keys are optional, by default all keys are required.
• rotate: boolean, defaults to false - treats rows as columns and columns as rows.
• unique: number, string, or array of strings. - which fields represent the unique key for a row. defaults to the first field.

csvparser's Issues

CSV parser logic is broken

The parser logic is not human readable and it doesn't work very well. It does not allow for newlines embedded in escaped fields. For more info: http://www.ietf.org/rfc/rfc4180.txt

Parse unix timestamps by default

It seems the built in date parser is pretty useless, we should replace it with one that we actually want to use.

This is the primary candidate:

module.exports = {
    test: function (value) {
        return !isNaN(new Date(value));
    },
    parse: function (value) {
        return Math.floor(Date.parse(value) / 1000);
    }
};