Hi all:

After an imprudent execution of a Rsync script, I deleted 50k files of my hardisk. Fortunately, I could recover them (file names, attributes, etc), but I lost the folder structure.

This is what I have now:

List of the files that has been deleted with their full path:

"DISCO ROJO/recuperado/Video/mpg/file000160.mpg"
"DISCO ROJO/recuperado/Video/mpg/file000159.mpg"
..

In the other hand.. (thanks to diskdrill) I got all files on a flat folder structure using random directory names:

./temp68554
./temp68555
./temp68556
./temp68557

Each temp folder match with old folder structure... it means that I.E.. the content of "./temp68557" is the content of the old "DISCO ROJO/recuperado/Video/mpg/

I'm finding ways to identify temp folder name with the old one.. and do a mv / cp script for that.

Any ideas?

Thanks

Does the tempnnnnn directories or their contents have ANY indication of the previous directory structure?

Thanks for your response. No, no correlation between tempnnnnn and old structure.

My best approach is:

1) create a matrix with all tempnnnn folder, and files inside. ( recursive ls of each tempnnnnn folder)
2) Create a second matrix with all original folders and files inside (I already extract old folders into a file.. so extract only filenames of the file for each old folder)
3) compare only directories with same number of files. (sort matrix entries based on the number of records for each folder)
4) Check that filenames are the same in both directories (loop until changes=0)
5) if #4 is ok, mark the source folder as "solved", > output the mv tempnnnn original structure command
6) if no jump to the next


Such a huge batch...

Looks horrible.

I should say also that filenames are not unique.. therefore I must match all filenames on each folder to be sure that the folder is the same one.

There is another one better:

1) Extract all filenames and full path on each (old and new one). Finding the last / give us the way to do it.
2) Find for all tempnnnnn folders (all files on each).. files on the old structure that match their filename. For each match, one counter add +1 on each counter (1 counter per domain on which filename match).. I.E... file 1.MP4 match on folder A > Counter A=A+1. If there are more coincidences (1.MP4 is not unique) Counter B=B+1.. etc.. In any case.. one counter must match at the end the total number of files of the folder. Than counter shows the correspondence tempnnnnnn <=> folder A
3) rename tempnnnnn to folder A etc etc etc.

Any tip?

Thanks