I don't know if any of your files will be there or not. And what I
say below only applies to text files, or ascii files that you can
read, to decide where they end. But you called it data so maybe it
is like that.
These files / directories are usually the result of Chkdsk / Scandisk
scans. It has always be extremely difficult to re-build data from
these.
When I hae files missing (which used to happen after a crash
sometimes) I found that chkdsk made one file for every file that was
missing. This was clear with text files and some ASCII data files
when I knew what the data looked like. The file itself was at the
beginning of the chkdsk file and after a certain point, there was just
garbage I guess whatever was in the cluster before the chkdsk wront a
file to it. Some editors have no limit on the size of hte file they
will edit, and with that, one could delete the part that was garbage
at the end, and rename the resulting file back to the original name,
and put it in the original directory. Since every file you find will
be at most only one file that you want, well, I don't know what your
results will be.
The best and by far the fastest way to look at dozens or hundreds of
files, most of which you won't want to edit, is by using TCC LE 10.0
or 11.0, which is now free at
www.jpsoft.com . It may be hard to find
it there, and if so write me, but before you do that google TCC LE
download That's how I found it the last time. It's right there on
the jpsoft site, and I found it there once, but the next time I
couldn't. It is free for individual use. Once you have installed
that, use the List command. All you do is press esc and it will go to
the next file, no matter how long it is, in an instant. No file
loading time. I used it a lot to look at chkdsk files.
I miss mainframes, where one can copy a designated number of bytes
from one file to another. Either at the start or even in the middle.
Isn't there a dos command or windows program that will do that? I've
looked and the only way seems to be an editor. But I don't think it
matters here because the method above works.
OTOH, some programas like newsreaders and email programs write loads
of posts or emails for example to the same file, I had a program with
a bug, and it didn't update that outbox file when it said it did, and
when there was a crash, every post I had written in that sesssion was
lost. In that case I had to remember some keywords from one of my
posts, and hunt through the whole harddrive for them, (Harddrives were
a lot smaller than, but slower) and there would be mulitiple
occurrences if I had saved any email or post more than once while
writing it, but when I eventually found the last version, it was like
a mother load, and I could pretty quickly copy out each one
separately, skipping the header information which was there, and
resend them or save them or whatever. For that I used the Norton
Utility diskedit, which ran in DOS. (There is some sort of disk
editor Norton includes that runs in windows iirc, but it can't do what
the dos version does.) Whether it runs in XP cmd, I don't know,
because the bug I mention was fixed years ago.
Now you have learned why you should backup every day, or more often,
or at least every time you get something you value.
