Fix batch upload id management

Happens when:

• Trying to attach 10 or more files to a document using the REST API.
• Using numerical ids starting from 1 for file ids (e.g. 1,2, 3 ... 10).

=> Content end up being mixed: the file name does not correspond to the actual stored blob when the batch upload is attached to the document.

Reason is ids are not read in the proper order, but that way:

• 0
• 1
• 10
• 2
• 3
• 4
• ...
• 9

Prefixing file ids by a 0 does not help. Workaround is to use larger numbers (ex, start from 1000, not 1).