couchDB: storing MB per document

Question

I am analyzing couchDB at the moment. Is it possible to storage MB worth of data per document? E.g. a JPEG image.

I understand I would need to encode (base64 or something) the said data in order to fit the JSON container.

Practical advice sought please.

Answer 1

As zed said in his comment the best way to do this is using attachments. The Wiki has a section on this: http://wiki.apache.org/couchdb/HTTP%5FDocument%5FAPI#Attachments

the basic idea is like so:

{
  "_id":"attachment_doc",
  "_attachments":
  {
    "foo.txt":
    {
      "content_type":"text\/plain",
      "data": "VGhpcyBpcyBhIGJhc2U2NCBlbmNvZGVkIHRleHQ="
    }
  }
}

You are correct that you should Base64 encode the attachments contents. You can have multiple attachments per document.

NOTE from the wiki: Please note that any base64 data you send has to be on a single line of characters, so pre-process your data to remove any carriage returns and newlines.

Answer 2

I have never tried big documents but I'm using documents with big attachments (JPEGs > 10 Mpix) and that works well.

The main issue is that which such huge database sizes replication tend to break in new and interesting ways every week.