BulkBadger

Takes a stream of Line Delimited JSON, e.g. from MongoDB Queries, PostgreSQL (> 9.2) Queries or from the filesystem and makes them suitable for bulk imports into other systems, most notably CouchDB, by batching the line delimited JSON.

stream.pipe(new BulkBadger({chunksize: 300}).pipe(outputStream)

Input:

{"rocko": "artischocko"}
{"zett": "zettmeister"}
{"mr": "mussie"}

Output:

{"docs": [
  {"rocko": "artischocko"},
  {"zett": "zettmeister"},
  {"mr": "mussie"}
]}

Output chunksize set to 2:

[
  {"docs":[{"rocko":"artischocko"},{"zett":"zettmeister"}]},
  {"docs":[{"mr":"mussie"}]}
]

Options

chunksize:        amount of docs in each chunk, default: 200

Examples

Use a regular JSON file from the fs as input

testjson.json:

[
  {"a": "b"},
  {"b": "c"},
  {"c": "d"}
]
 

var BulkBadger = require('bulkbadger')
 
var fs = require('fs')
var JSONStream = require('JSONStream')
 
fs
  .createReadStream('./testjson.json')
  .pipe(JSONStream.parse('*'))
  .pipe(new BulkBadger({chunksize: 2}))
  .pipe(JSONStream.stringify())
  .pipe(process.stdout)
 

Use a CSV file as input

var BulkBadger = require('bulkbadger')
 
var parse = require('csv-parse')
var fs = require('fs')
var transform = require('stream-transform')
var JSONStream = require('JSONStream')
 
 
var parser = parse({comment: '#', delimiter: ':'})
var input = fs.createReadStream('/etc/passwd')
 
 
var transformer = transform(function (record, cb) {
 
  var username = record[0]
  var pw = record[1]
  var uid = record[2]
  var gid = record[3]
  var comment = record[4]
  var home = record[5]
  var shell = record[6]
 
  cb(null, {
    id: username,
    pw: pw,
    uid: uid,
    gid: gid,
    comment: comment,
    home: home,
    shell: shell
  })
 
})
 
input
  .pipe(parser)
  .pipe(transformer)
  .pipe(new BulkBadger({chunksize: 2}))
  .pipe(JSONStream.stringify())
  .pipe(process.stdout)
 

Stream from MongoDB into CouchDB

var MongoClient = require('mongodb').MongoClient
var BulkBadger = require('bulkbadger')
var CouchBulkImporter = require('./index.js').CouchBulkImporter
 
var url = 'mongodb://localhost:27017/test'
// Use connect method to connect to the Server
MongoClient.connect(url, function (err, db) {
  console.log('Connected correctly to server')
  var col = db.collection('restaurants')
  var stream = col.find({}, {})
  stream
    .pipe(new BulkBadger({chunksize: 300}))
    .pipe(new CouchBulkImporter({
      url: 'http://tester:testerpass@localhost:5984/restaurant'
    })).on('error', function (e) {
      console.log('Oh noes!')
      console.log(e)
    })
 
  stream.on('error', function (e) {
    console.log('Oh noes!')
    console.log(e)
  })
  stream.on('end', function () {
    console.log('migration finished')
    db.close()
  })
})

Use Line-Delimited JSON as input

ldjson.json:

{"rocko": "artischocko"}
{"zett": "zettmeister"}
{"mr": "mussie"}

var fs = require('fs')
var JSONStream = require('JSONStream')
 
var BulkBadger = require('bulkbadger')
 
fs
  .createReadStream(__dirname + '/ldjson.json')
  .pipe(JSONStream.parse())
  .pipe(new BulkBadger({chunksize: 2}))
  .pipe(JSONStream.stringify())
  .pipe(process.stdout)