Question

I’m working on a Python/MongoDB project in both my computer at home and my laptop.
The schema in document stores, naturally, is best represented by the data itself – and that’s why I want to distribute my test data over Mercurial, together with the code itself.

Would the best way be to simply dump the BSONs in a file and add it to the mercurial repository?

Answer 1

Dumping BSON and putting it into a VCS would not make much sense, since it’s binary file and can’t be viewed easily.

You can export a collection to a JSON, by using mongoexport tool. You can even pass it a query filter to limit number of exported documents.

Here’s an example (reformatted for readability):

sergio@soviet-russia$ mongoexport -d test -c geo \
sergio@soviet-russia$    -q '{"_id":  ObjectId("4efa5f7d8840e680c850cd94") }'
connected to: 127.0.0.1
{ "_id" : { "$oid" : "4efa5f7d8840e680c850cd94" }, 
  "longg" : [ { "start" : 322815488, "end" : 322817535 }, 
              { "start" : 822815488, "end" : 822817535 } ], 
  "m" : "Cracow", 
  "postal" : 55050, 
  "lat" : [ "XX.89XXX", "XX.74XXX" ] }
exported 1 records