EXCEL file corruption

D

Dean

I have asked this question on another forum but there was no magic answers for me, so I am trying it here - I am desperate, as the file in question represents months of work!

I have a 10 MB file I have been expanding/refining over the last few months. Last week, in the middle of mere cosmetic upgrading, it crashed. Apparently, EXCEL now goes into a repair mode when that happens. Here is the message I got:

Microsoft Excel File Repair Log. Errors were detected in file 'C:\Documents and Settings\Dean\Application Data\Microsoft\Excel\PM 03-02 - 4 (version 2).xls'. The following is a list of repairs: Lost Data Validation information in one or more sheets.Damage to the file was so extensive that repairs were not possible. Excel attempted to recover your formulas and values, but some data may have been lost or corrupted.
The "repaired" file looked like (and acts like) a version I had (fortunately) saved just seven minutes earlier, except that all cosmetics had been wiped out, such as fonts, data validation choices, etc - not to mention extensive cell comments. If I change a few inputs, both files still produce identical results. Unfortunately, if I were to use the repaired file and redo the cosmetics, it would be many, many hours of work. Also, the error message above suggests it might still have corruption in it, so I'm not sure what I would gain.

Over the years, this happened to me twice with two different files. The first time, my file needed to be rebuilt from scratch. By the second time, EXCEL had some sort of manual algorithm that peeled back the onion until there was no longer any evidence of corruption. As it turned out, the thing that seemed to fix it was the removal of some sort of cell formatting. Now, EXCEL tries to repair the file automatically. In this case, I guess it is telling me it cannot guarantee success.

Given that fact, it seems silly to think of using the repaired file, when I have a working file just seven minutes earlier. However, regardless of which file I would use, there is an issue of how such corruption takes hold. Is it a one time spontaneous event, and if so, is it likely to recur? In the past two times, it did not recur, which made it seem pretty random! If it is not a one time event, but rather some sort of 'cumulative corruption' (the straw broke the camel's back), what confidence can I have that the results are correct? Also, how comfortable can I be in my strong belief that the corruption is limited to cosmetics, and that the equations are not compromised?

Is there a max EXCEL file size that is too big to be stable? Long ago, when this first happened, someone at Microsoft, off the record, gave me a file size number but I forget it. I recall he said that EXCEL was not a programming language and people were asking too much of it, sometimes, vis-a-vis monstrous, interconnected logic in files, etc. The fact that EXCEL can now do so much it couldn't do a decade ago, it doesn't seem that Microsoft is concerned, either that, or the architecture is so much improved.

I know these are hard questions and the last guy basically told me to use my own judgment, but if there is any research/info/experience out there, I would love to hear of it.

Thanks so much!

Dean
 
N

NickHK

Dean,
People have mentioned success with opening the corrupt Excel file in
OpenOffice, then resaving.

NickHK

I have asked this question on another forum but there was no magic answers
for me, so I am trying it here - I am desperate, as the file in question
represents months of work!

I have a 10 MB file I have been expanding/refining over the last few months.
Last week, in the middle of mere cosmetic upgrading, it crashed.
Apparently, EXCEL now goes into a repair mode when that happens. Here is
the message I got:

Microsoft Excel File Repair Log. Errors were detected in file 'C:\Documents
and Settings\Dean\Application Data\Microsoft\Excel\PM 03-02 - 4 (version
2).xls'. The following is a list of repairs: Lost Data Validation
information in one or more sheets.Damage to the file was so extensive that
repairs were not possible. Excel attempted to recover your formulas and
values, but some data may have been lost or corrupted.
The "repaired" file looked like (and acts like) a version I had
(fortunately) saved just seven minutes earlier, except that all cosmetics
had been wiped out, such as fonts, data validation choices, etc - not to
mention extensive cell comments. If I change a few inputs, both files still
produce identical results. Unfortunately, if I were to use the repaired
file and redo the cosmetics, it would be many, many hours of work. Also,
the error message above suggests it might still have corruption in it, so
I'm not sure what I would gain.

Over the years, this happened to me twice with two different files. The
first time, my file needed to be rebuilt from scratch. By the second time,
EXCEL had some sort of manual algorithm that peeled back the onion until
there was no longer any evidence of corruption. As it turned out, the thing
that seemed to fix it was the removal of some sort of cell formatting. Now,
EXCEL tries to repair the file automatically. In this case, I guess it is
telling me it cannot guarantee success.

Given that fact, it seems silly to think of using the repaired file, when I
have a working file just seven minutes earlier. However, regardless of
which file I would use, there is an issue of how such corruption takes hold.
Is it a one time spontaneous event, and if so, is it likely to recur? In
the past two times, it did not recur, which made it seem pretty random! If
it is not a one time event, but rather some sort of 'cumulative corruption'
(the straw broke the camel's back), what confidence can I have that the
results are correct? Also, how comfortable can I be in my strong belief
that the corruption is limited to cosmetics, and that the equations are not
compromised?

Is there a max EXCEL file size that is too big to be stable? Long ago, when
this first happened, someone at Microsoft, off the record, gave me a file
size number but I forget it. I recall he said that EXCEL was not a
programming language and people were asking too much of it, sometimes,
vis-a-vis monstrous, interconnected logic in files, etc. The fact that
EXCEL can now do so much it couldn't do a decade ago, it doesn't seem that
Microsoft is concerned, either that, or the architecture is so much
improved.

I know these are hard questions and the last guy basically told me to use my
own judgment, but if there is any research/info/experience out there, I
would love to hear of it.

Thanks so much!

Dean
 
D

Dean

Unless there is a remnant saved somewhere, before the automatic attempt at
repair, I don't think I had the choice to save or open the corrupt file. It
automatically tried to repair it and, in doing so, it removed all
formatting.

As I've said, I have a near final version, saved 7 minutes before the crash
and it 'seems' to work fine. So the problem is not recovering the file.
It's trying to understand what led to its corruption and whether some of
that may exist in the prior version(s). I am concerned about latent
corruption that could have been growing in the file for a while and could be
(maybe even was, before the crash) compromising the results.

I'm really trying to understand what usually corrupts EXCEL files (e.g.,
only cosmetics, not hairy equations) and what the impact might be (perhaps,
wiping out cosmetics only, not equations and/or data).

Thanks for your response. Please don't give up on me yet, everyone!

Dean
 
P

Peter T

Hi Dean,

A possibility might be if your wb has many unique cell formats, Red + bold +
font.size 12 + right border is a unique format. In your large wb, have
sheets been copied from other wb's. If the Normal styles in source &
destination wb's are not same a new bunch of unique formats can get added.
The limit is about 3,000 which sounds a lot but it's surprisingly easy to
reach. Normally a warning message should advise but it doesn't always and
then it's too late.

Regards,
Peter T
 
D

Dean

I don't recall how I created some of this but it may have legacy to other
workbooks created by another author at a different company. I am also
starting to become aware that it may be best to do formats in blocks, as
much as possible and I didn't do that, in many cases, especially having to
do with borders, where I often fixed one or two cells here and there.

This is starting to sound like a credible explanation . If the corruption
is of this type, is it pretty obvious that only the formats will be
corrupted (a knowable disaster I could live with), not the equations (an
unknowable disaster I could not live with)?

Thanks so much!
Dean
 
P

Peter T

Hi Dean
This is starting to sound like a credible explanation .

Might be credible but no guarantee it's correct!
If the corruption
is of this type, is it pretty obvious that only the formats will be
corrupted

Could be too late by that stage, ie not possible to re-open the file, keep
Nick's suggestion about OpenOffice in reserve.

Might be worth deleting custom styles if you have loads. Also apply
"consistent" formatting throughout. A reduction in file size after doing
that would be indicative.

Regards,
Peter T
 
D

Dean

<< Might be credible but no guarantee it's correct!>

Yes, this is a good point, Peter, since:

In my case, I can open a file that I saved 7 minutes earlier and I know
that, in those 7 minutes, I just did a lot of minor cell formatting. When I
went back and got that working file (after the crash), I tried to re-enact
those 7 minutes of changes (might have been done slightly differently), plus
some further stuff, and there was no problem thereafter, nor now. If I had
been on a hairy edge of too much formatting, I should be seeing similar
symptoms now, right? I don't. Or is it possible that, if you are doing a
lot of copying and pasting, EXCEL can just get confused (for example, if you
hit escape, or your computer is overloaded with other programs running) and
cause the file to self corrupt? Could this have anything to do with the
EXCEL program, rather than the file, being the source of that behavior, or
would EXCEL not have launched into the 'repair the file" mode, if such were
the case?

Just to belabor the point, I assume hairy equations, which I have many of,
take up more MB than oodles of data (which I have little of) - is that
right? Still, though a reduction in file size (after cleaning up
formatting) might be a good sign, I am getting the hints that my little 10
MB file is peanuts compared to many files used 'out there' with apparent
success. Is that true?

Nobody has addressed the automatic repair cycle EXCEL went into, before
delivering me a file that was devoid of all formatting. Is this the
expected mode if a file fails while you are in the middle of working on it?
Is this different from the corruption that most people see where they have
saved a file, but then can't even open it. My scenario did not happen in
that latter manner.

I have done pretty extensive comparisons between the repaired file and the
7-minute older file and the results are always the same, for a variety of
scenarios, a virtual impossibility if EXCEL's repair actually did compromise
the computational integrity. I suppose both files could have the same
compromised integrity (which would worry me) but, if so, it could have been
there months ago. What do you think of that possibility?

Thanks!
Dean
 
P

Peter T

Hi Dean,

You've asked a lot of reasonable questions but afraid I can't answer any!
Except perhaps to say I agree that formulas tend to consume more file size
than data.

I also agree with your un-scientific remark about Excel sometimes getting
confused. The only experience I have of corrupted files is after repeatedly
with code applying & deleting many formats, though at all times I was under
the rough 3,000 limit and didn't get a warning. Hence my suggestion to look
at formats. But it's certainly not the only reason for corruption. In my
case I got the distinct impression Excel doesn't always catch up with
changes, retaining extinct stuff somewhere to cause problems later.

I've not read a definitive list of things that cause corruption, but many
like this -
Q. Help my file is corrupted, I don't know why, what can I do?
A. Try OpenOffice or open in a different version of Excel.

With your large wb of many formulas might be worth ensuring they are
arranged efficiently. Charles Williams has good info on his site
www.decisionmodels.com

Backup !!

Regards,
Peter T

PS - Another possibility - temporary files not deleted after some crash.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top