The job of finding/verifying the required files in this huge debris of recovered 'junk' is left to the user of course. I think it makes the guess based on the 'header' of the file-blocks that are being recovered, and can easily get confused, assigning wrong extension names to the recovered files/fragments.Īlso, most of the recovered 'files' may be corrupt as they are only fragments. If your are not sure about that, your files or parts of them may reside anywhere on the disk depending on the availability of free space at the the time of their creation, regardless of 'When' they were created/copied to the disk - thus the need to read entire disk to make sure no files/parts are missing during recovery.Īnd Yes, photorec can recover files by their types, although it is not perfect in guessing the filetypes that are being recovered. This will work only if you are sure about the area on which your files needing recovery are located. Haven't done it myself, but I think you can create a partial image of your partition with ddrescue (gddrescue, telling it 'Bytes to SKIP' + 'Size to copy' - see 'man ddrescue'), then run photorec on that partial disk image.

Testdisk