@iwsio/json-csv-node
This package extends @iwsio/json-csv-core specifically targeting Node.JS to support streaming. It's all the same code as before, just moved around. You can read more about it on my blog.
Usage
Buffered (Converts to CSV with an in-memory data source)
const { toCsv } = require('@iwsio/json-csv-node')
const csv = await toCsv(data, options) // toCsv returns Promise
// optionally, you can use a callback
toCsv(data, options, function(err, csv) {...}))
Streaming (Converts a "row at a time" from a stream source)
When using the streaming API, you can pipe data to it in object mode.
const { toCsvStream } = require('@iwsio/json-csv-node')
const things = [
{name: 'thing1', age: 20},
{name: 'thing2', age: 30},
{name: 'thing3', age: 45}
]
Readable.from(things) // <readable source in object mode>
.pipe(toCsvStream(options)) // transforms to string and emits lines
.pipe(process.stdout) // anything Writable
})
NOTE: All the aliases from previous versions remain intact.
I renamed this API mostly for cosmetics and to cleanup named exports for ES Modules. It made more sense to me to import { toCSV } from 'json-csv'
rather than importing the default and calling its prop. (I've been doing a lot of Typescript lately). But all the original versions are kept intact and tests are in place to assert they work.
const jsonCsv = require('@iwsio/json-csv-node')
jsonCsv.toCsv // buffered, returns promise, callback optional
jsonCsv.toCsvStream // stream, returns stream, callback optional
jsonCsv.buffered // alias to toCsv (buffered)
jsonCsv.stream // alias to toCsvStream (streamed)
jsonCsv.csv // alias to toCsvStream (streamed)
jsonCsv.csvBuffered // alias to toCsv (buffered)
Options
{
// field definitions for CSV export
fields :
[
{
// required: field name for source value
name: 'string',
// optional: column label for CSV header
label: 'string',
// optional: transform value before exporting
transform: function(value) { return value; }
}
],
// Other default options:
fieldSeparator: ",",
ignoreHeader: false
}
Advanced Example
Here, you can see we're using a deeper set of objects for our source data, and we're using dot notation in the field definitions like: contact.name
for the contact name.
const items = [
{
downloaded: false,
contact: {
company: 'Widgets, LLC',
name: 'John Doe',
email: 'john@widgets.somewhere',
},
registration: {
year: 2013,
level: 3,
},
},
{
downloaded: true,
contact: {
company: 'Sprockets, LLC',
name: 'Jane Doe',
email: 'jane@sprockets.somewhere',
},
registration: {
year: 2013,
level: 2,
},
},
]
const options = {
fields: [
{
name: 'contact.company', // uses dot notation
label: 'Company',
},
{
name: 'contact.name',
label: 'Name',
},
{
name: 'contact.email',
label: 'Email',
},
{
name: 'downloaded',
label: "Downloaded",
transform: (v) => v ? 'downloaded' : 'pending',
},
{
name: 'registration.year',
label: 'Year',
},
{
name: 'registration.level',
label: 'Level',
transform: (v) => {
switch (v) {
case 1: return 'Test 1'
case 2: return 'Test 2'
default: return 'Unknown'
}
},
},
],
}
(async () => {
let result = await toCsv(items, options)
console.log(result)
})()
Output
Company,Name,Email,Downloaded,Year,Level
"Widgets, LLC",John Doe,john@widgets.somewhere,pending,2013,Unknown
"Sprockets, LLC",Jane Doe,jane@sprockets.somewhere,downloaded,2013,Test 2