my situation is this:
I have similar objects that differ in certain fields.
Example:
object1:
- name:'Albert'
- home:'London'
- email:'albert@corp.co.uk'
object2:
- name:'Jennifer'
- home:'Berlin'
- tel:'00492212232'
object3:
- name:'James'
- data:BIG_CHUNK_OF_BINARY_DATA
There is no fixed scheme which can be used in the definition of a RDBMS table. The definition of the objects is flexible with the possibility to add a custom field for a single object.
BIG_CHUNK_OF_BINARY_DATA means binary data in the size range from several bytes to several tens of megabytes. It should be possible to manage more than 100 objects of this kind with an overall mass of binary data in the lower gigabytes (1-5GB).
Now I’m searching for a database/format in which these objects can be stored efficiently. A search in the textual fields should be possible, of course. Perhaps a combination of JSON and a NoSQL-db is suitable? Or are there better solutions?
The target platforms would be Windows (mandatory) and Android (desireable).
Greets
Martin
I suggest looking at Couchbase Single Server (CouchDb). It holds a bunch of JSON documents in a schema-less structure. Structure is created through the use of ‘Views’ or indexes. They have a version running on Android too, although this is still in early development.