Hi Fernando,
In addition to David’s reply, for most repositories, the size of all the revision files is much smaller than the content, but they provide useful data to see e.g. when embargos have been lifted, or what changes there were in the item data when it was saved. Combined with the data in the ‘history’ dataset, you can see who changed a record, what was changed, and when.

They shouldn’t cause any problems on your system.

If you want to get from the data you have, to the specific file, run the following queries in your database – based on the /id/file/[number] – 136799 in your example:
SELECT * FROM file WHERE fileid = 136799;

From the results, the datasetid should be ‘history’, and there should be a number in the ‘objectid’ field e.g. 12345.

If you select that ‘objectid’ from the history table, it will give you details of what caused the revision to the data:

SELECT * FROM history WHERE historyid = 12345;

The columns from history table should include datasetid (hopefully ‘eprint’), the objectid (the eprint ID – 6789 as an example below) and the revision number e.g. 10.

This revision number is the filename on-disk, within the item’s document folder: archives/ARCHIVEID/documents/disk0/00/00/67/89/revisions/10.xml

Each eprint should have a ‘revisions’ folder with numbered XML files in it.

I think very old versions of EPrints (v2, maybe v3.1?) stored these revision files differently, so if the repository existed before v3.3.12, early eprints may have other styles of revision files.

Cheers,

John