Storing some small (under 1MB) files with MongoDB in NodeJS WITHOUT GridFS


Question

I run a website that runs on a backend of nodeJS + mongoDB. Right now, I'm implementing a system to store some icons (small image files) that will need to be in the database.

From my understanding, it makes more sense NOT to use GridFS, since that seems to be tailored for either large files or large numbers of files. Since every file that I need to save will be well under the BSON maximum file size, I should be able to save them directly into a regular document.

I have 2 questions:

1) Is my reasoning correct? Is it ok to save image files within a regular mongo collection, as opposed to with GridFS? Is there anything I'm not considering here that I should be?

2) If my thought process is sound, how do I go about doing this? Could I do something like the following:

//assume 'things' is a mongoDB collection created properly using node-mongodb-driver

fs.readFile(pathToIconImage, function(err,image){
  things.insert({'image':image}, function(err,doc){
    if(err) console.log('you have an error! ' + err);
  });
});

I'm guessing that there's probably a better way to do this, since mongoDB uses BSON and here I'm trying to save a file in JSON before I send it off to the database. I also don't know if this code will work (haven't tried it).

UPDATE - New Question

If I have a document within a collection that has three pieces of information saved: 1) a name, 2) a date, and 3) an image file (the above icon), and I want to send this document to a client in order to display all three, would this be possible? If not, I guess I'd need to use GridFS and save the fileID in place of the image itself. Thoughts/suggestions?

Best, and thanks for any responses,
Sami

1
22
7/11/2012 10:28:52 PM

Accepted Answer

If your images truly are small enough to not be a problem with document size and you don't mind a little amount of extra processing, then it's probably fine to just store it directly in your collection. To do that you'll want to base64 encode the image, then store it using mongo's BinData type. As I understand it, that will then save it as a BSON bit array, not actually store the base64 string, so the size won't grow larger than your original binary image.

It will display in json queries as a base64 string, which you can use to get the binary image back.

25
7/11/2012 10:50:37 PM

I have been looking for the same thing. I know this post is old , but perhaps i can help someone out there.

var fs = require('fs');
var mongo = require('mongodb').MongoClient;
var Binary = require('mongodb').Binary;

var archivobin = fs.readFileSync("vc.exe"); 
// print it out so you can check that the file is loaded correctly
console.log("Loading file");
console.log(archivobin);

var invoice = {};
invoice.bin = Binary(archivobin);

console.log("largo de invoice.bin= "+ invoice.bin.length());
// set an ID for the document for easy retrieval
invoice._id = 12345; 
  mongo.connect('mongodb://localhost:27017/nogrid', function(err, db) {
  if(err) console.log(err);
     db.collection('invoices').insert(invoice, function(err, doc){
    if(err) console.log(err);
  // check the inserted document
    console.log("Inserting file");
    console.log(doc);

    db.collection('invoices').findOne({_id : 12345}, function(err, doc){
      if (err) {
        console.error(err);
        }

      fs.writeFile('vcout.exe', doc.bin.buffer, function(err){
          if (err) throw err;
          console.log('Sucessfully saved!');
    });

    });
  });
});

Licensed under: CC-BY-SA with attribution
Not affiliated with: Stack Overflow
Icon